Detecting underreporters of abortions and miscarriages in the national study of family growth, 2011-2015.

This paper draws on individual-level data from the National Study of Family Growth (NSFG) to identify likely underreporters of abortion and miscarriage and examine their characteristics. The NSFG asks about abortion and miscarriage twice, once in the computer-assisted personal interviewing (CAPI) part of the questionnaire and the other in the audio computer-assisted self-interviewing (ACASI) part. We used two different methods to identify likely underreporters of abortion and miscarriage: direct comparison of answers obtained from CAPI and ACASI and latent class models. The two methods produce very similar results. Although miscarriages are just as prone to underreporting as abortions, characteristics of women underreporting abortion differ somewhat from those misreporting miscarriages. Underreporters of abortions tended to be older, poorer, less likely to be Hispanic or Black, and more likely to have no religion. They also reported more traditional attitudes toward sexual behavior. By contrast, underreporters of miscarriage also tended to be older, poorer, and more likely to be Hispanic or Black, but were also more likely to have children in the household, had fewer pregnancies, and held less traditional attitudes toward marriage.

Introduction

Background

Survey respondents overreport socially desirable behaviors, like voting and donating money to charity, and underreport socially undesirable behaviors, like using illicit drugs or drinking too much (see [1] for a review). One behavior that generally is underreported in surveys is having had an abortion. In a series of studies, researchers have compared estimates of the number of abortions in the United States based on reports from the National Survey of Family Growth (NSFG) with estimates based on surveys of U.S. abortion providers. These studies have consistently shown that the NSFG estimates are too low—the survey respondents apparently report only about half of their abortions [2-6]. For example, a study [4] estimates that in the 2002 NSFG, respondents reported only 47 percent of their abortions. Similarly, Tourangeau and his colleagues used a sample that included women who were known to have had abortions and found that fewer than 75 percent of them reported ever having had an abortion and only about half reported an abortion during the period in which they were known to have had one [7]. The problem of underreporting of abortions is not limited to the NSFG; abortions are also underreported in other national surveys in the U.S., including the 1997 National Longitudinal Survey of Youth and the Add Health Study, and in other countries [5, 8–11].

Less well studied is the underreporting about miscarriages. Even though both are pregnancy loss, miscarriage and abortion are differentially sensitive. Abortion is more stigmatized than miscarriage. One study [12] found about one in three respondents kept abortion as a secret whereas only 7% did so with miscarriage. In addition, of those who avoided telling about their abortion, 36% did so to avoid stigma. By contrast, less than 3% of respondents were concerned about stigma when they did not report their miscarriage. A second study compared [13] perception of abortion stigma to miscarriage stigma perception and found that the mean score on the abortion stigma perception is higher than that on the miscarriage stigma perception. Another study [14] further revealed that respondents erroneously perceived miscarriage as a rare complication of pregnancy. Those with a miscarriage felt that they lost a child and did something wrong, and felt guilty, alone, and ashamed [13, 14]. Abortions were sometimes misreported as miscarriage [8, 10]. There is limited evidence that miscarriage is also prone to misreporting [4, 15]. For instance, NSFG respondents were found to underreport miscarriage in the CAPI portion of the interview as often as they underreported abortion [15].

It is important to identify women respondents who are most prone to underreport abortions and miscarriages. A review of the literature on social desirability bias in survey reports argues that survey respondents often “edit” their answers prior to reporting them in order to avoid embarrassing themselves [1]. The review identifies several general things that affect whether survey reports will be subjected to such editing [1]. Respondents are more likely to misreport when an interviewer administers the questions than when the questions are self-administered; respondents in the socially undesirable category (e.g., non-voters) are much more likely to misreport than those in the desirable category (voters); and those who personally subscribe to the norms that make a given behavior socially desirable or undesirable are more likely to misreport than those who reject those norms. It is more embarrassing to admit something to an interviewer than to a computer and to have broken a norm than not to have broken one, especially when the norm is one that the respondent subscribes to.

We believe that some of the same factors that affect reports about other sensitive topics affect survey reports about abortion and miscarriage as well. For example, respondents are more likely to admit to having had an abortion when the questions are self-administered than when they are administered by an interviewer [8, 16, 17], although one exception is reported in [7]. Based on these findings, the NSFG asks two sets of questions about abortion, one set that is administered by the interviewers and a second set that is self-administered. The self-administered questions consistently elicit more reported abortions than the interviewer-administered ones [18].

This paper has three goals. The first is to evaluate the utility of the latent class analysis (LCA) approach in identifying underreporters of abortion and miscarriage. It is usually impossible to identify which respondents misreported from the survey data alone; ideally, medical records are needed for comparisons to survey data to properly identify underreporters. With NSFG, we can use answers from CAPI and ACASI to identify underreporters of abortion and miscarriage—that is, those who reported abortion in only one of the two modes. The weakness of this method is the failure to identify women who don’t report abortion (or miscarriage) in either mode. LCA is a promising alternative because, in principle, it has the potential to identify those women who didn’t report their abortion in either mode without needing medical records. We will replicate and extend the findings of an earlier analysis [18], using a larger and more recent data set and examining a wider range of LCA models on two pregnancy outcomes. Then, to evaluate the utility of LCA models, we compare the results from LCA models to the results that use only survey data to identify likely underreporters.

The second goal is to try to understand factors affecting underreporting behavior of respondents by identifying the characteristics of respondents most prone to underreport abortion and miscarriage. Earlier work [4, 5] to characterize women underreporting abortion were conducted at the aggregate level and found that older women, married women, poorer women, college graduates, and Catholics are more likely to underreport abortions than their younger, unmarried, wealthier, less educated, and non-Catholic counterparts. We hope to replicate these findings and to extend them by using individual-level data and by examining additional characteristics of women associated with underreporting abortions. In particular, we test the idea that women from traditional cultural backgrounds, where disapproval of abortion is strong, are most prone to underreport their abortions.

The third goal is to investigate underreporting of miscarriages in the NSFG. We will compare the characteristics of women who underreport their abortions with those of women who underreport miscarriages to understand the differential sensitivity associated with abortion and miscarriage. We will examine whether or not there is a common set of characteristics that contribute to underreporting of both abortion and miscarriage.

Methods

Our study is based on data from the National Survey of Family Growth (NSFG), a survey sponsored by the National Center for Health Statistics and, since 2002, carried out by the University of Michigan’s Survey Research Center. The NSFG has been done periodically since 1973 and moved to a continuous design (with data collected every quarter) in 2006. We analyze data from the 2011 to 2015 NSFG; the samples during this period include data from 11,300 women respondents (and 9,321 men); our analysis is based solely on the women.

Sample design

The target population for the 2011–2015 NSFG is the noninstitutionalized population 15–44 years old, whose usual place of residence is the 50 United States and the District of Columbia. It excludes people living in institutions, such as prisons or military bases.

To represent this population, the NSFG used a stratified five-stage area probability sample. The first stage of sample selection consisted of the selection of primary sampling units (PSUs); each PSU was a metropolitan area, a single county, or a group of counties. Prior to selection, all the PSUs were grouped into strata, based on census region and division, PSU size, and PSU Metropolitan Statistical Areas (MSA) / Non-MSA status. At the next stage, blocks or groups of adjoining blocks—second stage units (SSUs)—were selected. Both the PSUs and SSUs were selected with probability proportionate to size, where the size measure gave a higher selection probability to areas where at least 10 percent of the population was Black or Hispanic. In the third stage, a list of the housing units (HUs) within each SSUs was compiled and sample of HUs was selected. Interviewers either updated an existing list of addresses (based on the U.S. Postal Service’s Delivery Sequence File) or created a list from scratch in SSUs where no list was available. The sample HUs were contacted and a short screening interview was administered. In units with eligible residents (that is, someone in the 15–44 age range), a fourth stage of selection was carried out that involved a random selection of one eligible person for the main interview. The within-household selection rates were set so that about 20 percent of all the interviews were with adolescent respondents (aged 15–19) and about 55 percent were with females. The final stage of sampling was carried out during the last two weeks of each quarter’s 12-week field period; at this point, a subsample of the remaining cases was selected for continued follow-up.

Data collection

Both the screener and most of the main NSFG interview were done via computer-assisted personal interviewing (CAPI); audio computer-assisted self-interviewing (ACASI) was used for the final section of the main interview, which contained the question items considered most sensitive (such as alcohol and drug use, involuntary sex, sexual disease, sexual orientation, and so on). The entire interview was programmed in the Blaise software (version 4.8). The ACASI portion of the interview featured text-to-speech—that is, a computer-generated voice—rather than a recorded human voice.

Under the NSFG’s design, new samples are released each quarter. Interviewers attempt to complete all their assigned cases during the first ten weeks of each quarter; then, during the final two weeks of the 12-week field period, any remaining nonresponding cases are subsampled and interviewers attempt to complete the cases retained in the subsample. During this final phase of data collection each quarter, the incentive offered to respondents is doubled from $40 to $80.

Key items

The questionnaire for female respondents consists of ten sections; the key abortion and miscarriage questions come in the second and tenth sections. In the second section, the respondent is asked whether she might be pregnant currently and how many times in total she has been pregnant. She is then asked about the outcome of each pregnancy (miscarriage, stillbirth, abortion, ectopic or tubal pregnancy, or live birth) and the month and year when the pregnancy ended. From these items, derived variables are constructed, indicating the number of abortions and miscarriages the respondent reported in the past five years under CAPI. At the beginning of the final section of the questionnaire, respondents are asked in separate questions (this time in ACASI), how many pregnancies they had in the last five years that ended in a live birth; a stillbirth, miscarriage, or tubal or ectopic pregnancy; or an abortion. The data from the second and the ACASI sections of the questionnaire are recoded into dummy variables denoting a report of at least one abortion (or miscarriage) in the last five years. They are then used as the main indicators for the LCA models. Table 1 shows the (unweighted) cross-tabulation of the key indicators of abortion and miscarriage reporting.

Table 1

Unweighted cross-tabulation of the ACASI and CAPI indicators of abortion and miscarriage.

		CAPI Indicator
		YES (Abortion Reported)	NO (No Abortion Reported)
ACASI Indicator	YES (Abortion Reported)	422	273
	NO (No Abortion Reported)	39	10544
		CAPI Indicator
		YES (Miscarriage Reported)	NO (No Miscarriage Reported)
ACASI Indicator	YES (Miscarriage Reported)	840	506
	NO (No Miscarriage Reported)	69	9863

Analytical strategy

We first used only survey reports to identify consistent reporters of abortion as those who reported abortion in both modes and likely underreporters of abortion as those who reported abortion in only one mode. As shown in Table 1, 422 women are classified as consistent reporters of abortion and 312 (= 39+273) women as likely underreporters of abortion. In the same manner, 840 women are classified as consistent reporters of miscarriage and 575 (= 69+506) women as likely underreporters of miscarriage. For the regression analyses, we assume that consistent reporters of abortion and miscarriage are truthful reporters of abortion and miscarriage.

Then, we adopted a three-step approach to apply LCA in the analysis. LCA models the relationships among a set of observed categorical variables (in this case, the CAPI and ACASI “indicators” of abortion/miscarriage) measuring one unobserved (that is, “latent”) categorical variable with two or more classes (in this case, a two-class latent abortion/miscarriage variable). The associations between the observed variables reflect the fact that the population consists of a set of mutually exclusive and exhaustive latent classes with different distributions on the observed variables. Within each of the latent classes, the observed variables are unrelated. It is this key assumption of “local independence” that allows inferences about the latent class variable [19].

In such a model, the probability of an observed response (μ) on question j depends on the conditional probability of observing that response given that the respondent is in latent class k, summed across all K of the latent classes. Given that the responses are independent of each other within each latent class, the probability of the vector of responses μ is: in which there are K latent classes, each with a “prevalence” (i.e., unconditional probability) of P(c = k). The model produces estimates of these unconditional probabilities—representing the relative sizes of each latent class—as well as of the conditional probabilities of each response within each latent class (P(μ|c = k)).

For a two-class LCA, three indicators are necessary for the model to be identified [19]. When there are only two available indicators, researchers can choose to impose various restrictions on the LCA model parameters to achieve identifiability. For example, the false positive probability might be assumed to be zero or the latent classes might be assumed to be the same size. (Additional examples are provided in [20]). However, such assumptions are often implausible. Another approach is to include a grouping variable (G = 1, 2,…, g) in the model that predicts membership in the latent class, as in the Hui-Walter model [21-24]. To achieve an identifiable model, the Hui-Walter model makes two assumptions about the grouping variable:

The prevalence rates differ by the level of the grouping variable (the unequal prevalence assumption); and

The false positive and false negative probabilities are the same in each level of the grouping variable (the equal error probabilities assumption). (False positive error arises when a respondent is assigned to the ‘did not have an abortion’ latent class but she reported having had one, whereas a false negative error occurs when a respondent is assigned to the ‘had an abortion’ latent class but did not report having one.)

In this paper, we first fit a two-class latent class model. Since there were only two indicators of abortion/miscarriage, we adopted the Hui-Walter approach and used marital status as a grouping variable to achieve an identifiable model. The model also included additional covariates, such as age (20 to 29 years old versus all others), poverty level (three classes); and race/ethnicity (Black or Hispanic women versus all others). In addition, we explored models that included a third indicator of abortion/miscarriage–whether the respondent reported in the CAPI section of the questionnaire ever being pregnant. We used PROC LCA in SAS to fit the models [25].

Next, we assigned each respondent to one of the two latent classes (women who had had an abortion in the last five years and those who had not) based on their posterior class membership probabilities. We used two different methods to make these assignments. In the first method, modal assignment, each respondent was assigned to the latent class to which she had the highest probability of belonging. In the second method, stochastic assignment, the respondent was randomly assigned to a latent class, with the assignment probability equal to her posterior class membership probability.

We then compared respondents’ reports of their abortion status to their predicted class membership. We divided the respondents into four groups: 1) those who were assigned to the class of women who had had an abortion in the last five years and reported it in both the CAPI and ACASI portions of the questionnaire; 2) those who were assigned to that class but did not report an abortion in at least one mode; 3) those who are assigned to class of women who had not had an abortion in the last five years, but reported an abortion in at least one mode; and 4) those who were assigned to the class who had not had an abortion in the last five years and did not report an abortion in either mode. The first group of respondents are truthful reporters of abortion (according to the model) whereas the second group are underreporters of an abortion because they failed to report having an abortion in either CAPI or ACASI. The third group, according to the model, represents those who overreport abortion and the fourth group, accurate reporters of no abortion.

In the third step, we fit logistic regression models to compare those who underreported their abortions in at least one mode (i.e., reporting group 2 above) to truthful reporters of abortions (i.e., reporting group 1) on a wide range of variables related to abortion. Predictors include age (as a continuous variable), marital status (a dummy variable contrasting “married or cohabitating” vs. all others), race and ethnicity (a dummy variable contrasting “Hispanic or Non-Hispanic Black women” vs. all others), total family income recoded into 15 income brackets, whether there were no children under the age of 18 in the household (= 1) or at least one (= 0), the number of pregnancies in the lifetime (as a continuous variable), the number of lifetime male sexual partners (as a continuous variable), whether the respondent currently reported no religion (= 1) or any religion (= 0), whether the respondent lived in a metropolitan area (= 1) or not (= 0), whether the respondent’ mother had a high school or less education (= 1) or more than high school (= 0), whether the respondent was born outside of USA (= 1) or not (= 0), and whether the respondent completed the ACASI interview in Spanish (= 1) or not (= 0).

We also created three scales for possible inclusion in the models. The first (“Risky substance use behaviors”) was a count of the number of risky substance use behaviors the respondent reported, including having smoked at least 100 cigarettes in her lifetime; having drunk beer and other alcoholic beverages; smoked marijuana; used cocaine, crack, Crystal or meth; and injected drugs other than prescriptions at least once or twice during the year. The second and third scales were based on a battery of attitudinal items on sex, divorce, and homosexuality. All of these items used a five-point agree-disagree response scale. We created a scale of traditional sexual attitudes, based on the respondent’s answers to eight of the items (e.g., Sexual relations between two adults of the same sex is all right). These eight items loaded highly (absolute value of .55 or higher) on the first component of an exploratory factor analysis of the items. [S1 Table] gives the exact wording of all eight items and their loadings on this scale. Higher scores on the index indicate more traditional attitudes. Our hypothesis was that the women who had more traditional attitudes would be more reluctant to report an abortion and thus more likely to be underreporters. The factor analysis also yielded a second factor, which we labelled attitudes toward marriage. This scale was based on answers to three of the items (e.g., Divorce is usually the best solution when a couple can’t seem to work out their marriage problems). Again, higher scores indicate more traditional attitudes. [S2 Table] gives the exact wordings for all three of these items and their loadings on this scale. We thought that woman with more negative attitudes toward marriage might be more likely to report their abortions.

The logistic regression models were run in SAS (PROC SURVEYLOGISTIC), accounting for the complex sample design (that is, the weights, stratification, and clustering).

We used the same analytic strategy to identify and predict underreporters of miscarriage.

Results

Predicted latent classes for abortion and miscarriage reporting

Table 2 displays the prevalence of the two latent abortion classes produced by the two-indicator LCA that drew only on the CAPI and ACASI responses and by the three-indicator LCA that included whether respondents reported they were ever pregnant (= 1) or not (= 0) as the third indicator. As noted earlier, we used both the deterministic (assignment to the more likely latent class) and stochastic approaches (assignment based on the predicted posterior probabilities). The results are quite consistent across the LCA models and class assignment approaches. From 6.2% (n = 703) to 6.6% (n = 741) of the women were assigned to the “had an abortion” class, which is a bit higher than the proportion of self-reported abortions in either CAPI or ACASI modes, as shown in Table 1. The weighted figures are similar. Not surprisingly, LCA models show that the CAPI indicator of abortion has, on average, a higher false negative rate than the ACASI indicator (37% vs. 8%); in other words, more women who were assigned to the “had an abortion” latent class reported not having an abortion in the CAPI mode than in the ACASI mode. This is consistent with past findings [18] that CAPI produces more underreports of having had an abortion than ACASI.

Table 2

Predicted latent abortion and miscarriage classes (unweighted).

	2-indicator LCA		3-indicator LCA
Latent Abortion Class	Modal Assignment	Random Assignment	Modal Assignment	Random Assignment
Had an abortion in last five years	734 (6.5%)	741 (6.6%)	715 (6.3%)	703 (6.2%)
Did not have an abortion in last five years	10,544 (93.5)	10,537 (93.4%)	10,563 (93.4%)	10,575 (93.8%)
Total	11,278 (100%)	11,278 (100%)	11,278 (100%)	11,278 (100%)
	2-indicator LCA		3-indicator LCA
Latent Miscarriage Class	Modal Assignment	Random Assignment	Modal Assignment	Random Assignment
Had a miscarriage in last five years	1,415 (12.6%)	1,379 (12.2%)	1,393 (12.3%)	1,395 (12.4%)
Did not have a miscarriage in last five years	9,863 (87.4%)	9,899 (87.8%)	9,885 (87.7%)	9,883 (87.7%)
Total	11,278 (100%)	11,278 (100%)	11,278 (100%)	11,278 (100%)

Results on the prevalence estimates of the two latent miscarriage classes are similar across LCA models and class assignment methods. From 12.2% (n = 1,379) to 12.6% (n = 1,415) of women respondents were assigned to the “had a miscarriage” latent class. This is also higher than the unweighted percent of women reporting having had a miscarriage in either mode of data collection. Similarly, the CAPI indicator of miscarriage also has a higher false negative rate (34%) than the ACASI indicator (7%).

Overall, the findings were similar whether the latent class model was based on the Hui-Walter assumptions or incorporated a third indicator, and whether the class membership was assigned in a deterministic or stochastic manner.

Underreporters versus truthful reporters of abortion

We compared respondents’ self-reports of abortion to the latent class assigned, placing respondents in one of the four reporting groups. As shown in Table 3, the results are again quite consistent across two- and three-indicator LCA models and across the two methods of assigning respondents to a latent class. Regardless of the LCA model or assignment approach, 422 respondents are classified as truthful reporters of abortion; they reported they had at least one abortion in the last five years in both CAPI and ACASI. Close to 40% of respondents in the latent class of women who had an abortion (n = 281 to 319) are classified as underreporters of abortion; they denied having had an abortion in the last five years in least one of the two modes. The weighted figures are similar; from 43 to 47 percent of women underreport their abortions in at least one mode. Recall Table 1, 422 women are classified as truthful reporters of abortion and 312 women as underreporters of abortion using only survey data. The different methods of identifying underreporters of abortion produce very similar results.

Table 3

Abortion reporting behaviors (unweighted).

	2-indicator LCA		3-indicator LCA
	Modal Assignment	Stochastic Assignment	Modal Assignment	Stochastic Assignment
Truthful reporters of abortion	422	422	422	422
Underreporters of abortion	312	319	293	281
Underreporters as percent of those classified as having had an abortion	42.5%	43.0%	41.0%	40.0%
Overreporters of abortion	0	18	19	52
Truthful reporters of no abortion	10,544	10,519	10,544	10,523

Note: Figures are unweighted.

Only a handful of respondents are classified as overreporters of abortion. The vast majority of respondents who reported no abortion in either mode were classified as truthful reporters of no abortion.

We ran logistic regression models predicting the respondent’s likelihood of being classified as an underreporter (group 2 in Table 3) rather than as a truthful reporter of an abortion in the past five years (group 1 in Table 3). Table 4 presents model results in logit scale whereas [S3 Table] provides odds ratios, 95% confidence interval, and the p-values. As shown in Table 4, three variables were significantly related to a woman’s likelihood to underreport her abortion across all five of the models. Older women were more likely to underreport an abortion. By contrast, women with no religion and with higher incomes were less likely to underreport an abortion. Traditional sexual attitudes were positively associated with the women’s likelihood to underreport an abortion; the association was statistically significant in four of the five models and was marginally significant in the remaining model. The number of risky substance use behaviors the respondent reported was negatively related to the women’s likelihood to underreport an abortion. However, this relationship was statistically significant at p < .05 for three models and marginally significant under two more (p < .10). Finally, Hispanic or Black women were significantly less likely to underreport an abortion in two of the five models. It is worth mentioning that the model results using LCA classification are highly consistent with the model results using the classification based solely on the survey data.

Table 4

Logistic regression coefficients in models predicting underreporting of abortion.

	2-indicator LCA, modal assignment		2-indicator LCA, random assignment		3-indicator LCA, modal assignment		3-indicator LCA, random assignment		Survey Data Only
Parameter	Estimate	SE	Estimate	SE	Estimate	SE	Estimate	SE	Estimate	SE
Intercept	2.06	0.57	2.39	0.57	1.76	0.55	1.87	0.55	2.06	0.57
Age (centered at mean)	0.07**	0.02	0.07*	0.02	0.09**	0.02	0.11***	0.02	0.07**	0.02
Married or Cohabitating	0.15	0.26	0.17	0.24	0.37	0.27	0.49	0.26	0.15	0.26
Hispanic or Black	-0.55*	0.27	-0.49	0.28	-0.37	0.29	-0.45	0.31	-0.55*	0.27
No Children in Household	-0.27	0.27	-0.31	0.27	-0.51	0.30	-0.62	0.29	-0.27	0.27
Number of Pregnancies	-0.10	0.07	-0.15	0.08	-0.08	0.07	-0.11	0.07	-0.10	0.07
Number of Life Partners	-0.01	0.01	-0.01	0.01	0.01	0.01	0.01	0.01	-0.01	0.01
No Religion	-0.78**	0.24	-0.92**	0.23	-0.56*	0.24	-0.49*	0.24	-0.78**	0.24
Total Income	-0.09**	0.03	-0.08*	0.03	-0.10**	0.03	-0.10**	0.03	-0.09**	0.03
Metropolitan Area	-0.51	0.28	-0.51	0.28	-0.54	0.28	-0.57	0.29	-0.51	0.28
Mother with High School Education or Less	0.01	0.25	0.06	0.25	0.03	0.29	0.09	0.29	0.01	0.25
Born outside USA	0.08	0.41	-0.04	0.43	-0.06	0.34	-0.21	0.38	0.08	0.41
Interview Language	0.48	0.70	0.43	0.70	0.30	0.69	0.23	0.73	0.48	0.70
Risky Substance Use Behaviors	-0.25	0.15	-0.35*	0.17	-0.30*	0.14	-0.37*	0.15	-0.25	0.15
Traditional Sexual Attitudes	0.39*	0.17	0.29	0.17	0.45*	0.18	0.41*	0.18	0.39*	0.17
Attitudes toward Marriage	0.04	0.16	0.01	0.17	-0.02	0.18	-0.07	0.19	0.04	0.16
n	696		702		678		665		696
Pseduo-R	0.1846		0.1867		0.2146		0.2301		0.1846

Note:

* p<0.05

** p<0.01

*** p<0.001

Underreporters versus truthful reporters of miscarriages

Although miscarriages are presumably less sensitive to report than abortions, they are not always reported accurately. Table 5 shows the proportion of women classified in each reporting group for miscarriage. Regardless of the LCA model or assignment approach, 840 respondents are classified as truthful reporters of miscarriage; they reported having at least one miscarriage in the last five years in both CAPI and ACASI. Close to 40% of respondents in the latent class of women who had a miscarriage (n = 539 to 575) are classified as underreporters of miscarriage; they denied having had a miscarriage in the last five years in least one of the two modes. The weighted figures are similar; from 35 to 36 percent of women underreport their miscarriages in at least one mode. Table 5 shows that miscarriages are almost as prone to underreporting as abortions, a finding consistent with the previous findings [15]. Overreporters of miscarriages are, according to the latent class models, very rare. This classification pattern is very similar to the classifications based on the survey data only (see Table 1).

Table 5

Miscarriages reporting behaviors (unweighted).

	2-indicator LCA		3-indicator LCA
	Modal Assignment	Stochastic Assignment	Modal Assignment	Stochastic Assignment
Truthful reporters of miscarriage	840	840	840	840
Underreporters of miscarriage	575	539	553	555
Underreporters as percent of those classified as having had a miscarriage	40.6%	39.1%	39.7%	39.8%
Overreporters of miscarriage	0	78	22	57
Truthful reporters of no miscarriage	9,863	9,821	9,863	9,826

To what extent is underreporting of miscarriages associated with the same variables as underreporting of abortion? Table 6 presents model results in logit scale whereas [S4 Table] provides odds ratios, 95% confidence interval, and the p-values. As shown in Table 6, some variables, such as age and income are significantly related to underreporting both abortions and miscarriages. The odds of underreporting increases with age, but is lower for women with higher incomes. Still, there are some noteworthy differences in the predictors of underreporting of the two outcomes. Hispanic or Black women are significantly more likely to underreport miscarriages than other women, but not abortions (where the sign for this variable is in the opposite direction). Women with no children and women with more pregnancies are significantly associated with less underreporting of miscarriage, but the number of children and the number of pregnancies had no association with underreporting of abortion. Furthermore, women holding more traditional attitudes toward marriage were less likely to underreport miscarriages. Perhaps the key differences across the models are that risky substance use behaviors and the traditional sexual attitudes are significantly related only to abortion underreporting.

Table 6

Logistic regression coefficients in models predicting underreporting of miscarriage.

	2-indicator LCA, Modal Assignment		2-Indicator LCA, Random Assignment		3-indicator LCA, Modal Assignment		3-indicator LCA, Random Assignment		Survey Data Only
Parameter	Estimate	SE	Estimate	SE	Estimate	SE	Estimate	SE	Estimate	SE
Intercept	0.64	0.38	0.69	0.40	0.41	0.38	0.52	0.39	0.64	0.38
Age (centered at mean)	0.06***	0.01	0.05**	0.02	0.07***	0.02	0.08***	0.02	0.06**	0.01
Married or Cohabitating	-0.46*	0.20	-0.29	0.21	-0.26	0.20	-0.18	0.21	-0.46*	0.20
Hispanic or Black	0.43*	0.18	0.53*	0.19	0.52**	0.18	0.54**	0.18	0.43*	0.18
No Children in Household	-0.69**	0.23	-0.61*	0.22	-0.91***	0.25	-1.04***	0.26	-0.69**	0.23
Number of Pregnancies	-0.21**	0.06	-0.26***	0.07	-0.17**	0.06	-0.23**	0.07	-0.21**	0.06
Number of Life Partners	-0.02	0.01	-0.02	0.01	-0.01	0.01	-0.01	0.01	-0.02	0.01
No Religion	-0.06	0.22	-0.05	0.22	-0.02	0.22	0.02	0.22	-0.06	0.22
Total Income	-0.06**	0.02	-0.06**	0.02	-0.07**	0.02	-0.08**	0.02	-0.06**	0.02
Metropolitan Area	0.29	0.18	0.23	0.19	0.26	0.18	0.27	0.19	0.29	0.18
Mother with High School Education or Less	0.05	0.18	0.14	0.18	0.02	0.18	0.13	0.18	0.05	0.18
Born outside USA	0.09	0.26	0.06	0.26	0.10	0.27	0.08	0.26	0.09	0.26
Interview Language	0.10	0.42	0.00	0.43	-0.04	0.43	-0.10	0.45	0.10	0.42
Risky Substance Use Behaviors	0.13	0.11	0.06	0.11	0.10	0.11	0.11	0.10	0.13	0.11
Traditional Sexual Attitudes	0.01	0.12	-0.05	0.12	-0.01	0.13	-0.06	0.13	0.01	0.12
Attitudes toward Marriage	-0.34*	0.13	-0.36*	0.13	-0.37*	0.13	-0.39*	0.13	-0.34*	0.13
n	1351		1316		1330		1330		1351
Pseduo-R	0.0992		0.0973		0.1051		0.1164		0.0992

Note:

* p<0.05

** p<0.01

*** p<0.001

Models results from LCA classifications are, again, very similar to results from classifications using survey data, with only two exceptions. The effects of marital status on miscarriage underreporting are statistically significant for only two of the four LCA models.

Discussion

Because of the social stigma associated with abortion, women are found to underreport abortions in surveys [e.g., 4, 5]. Although the use of ACASI reduces the extent of underreporting, it does not eliminate it. It is, therefore, important to know which respondents are more likely to underreport their abortions. We used latent class analysis (LCA) to compare underreporters of abortions to truthful reporters. The key advantage of LCA lies in the fact that it does not require an error-free indicator of abortion. In this paper, we used LCA to predict the probabilities that respondents from the NSFG fell into one of two latent classes—woman who had had an abortion in the last five years and those who had not. We then used these probabilities to assign women to a latent class. We ran LCA models that used two or three indicators and we used two different methods to assign respondents to a latent class. The results are quite consistent across the different models and methods of assignment. About 7 percent of the women are assigned to the latent class of women who had had an abortion in the last five years and, among them, about 40 percent did not report having had an abortion in at least one mode. This level of underreporting is consistent with other research, though lower than the rates of underreporting found in comparisons to the Guttmacher Institute abortion providers census [4, 5].

Among those classified as having had an abortion, we distinguished underreporters from truthful reporters of abortion, based on their survey answers. We examined their demographic characteristics, fertility characteristics, and other characteristics (whether the respondent lived in a metropolitan area, whether the respondent’s mother had high school or less education, whether the respondent had a religion, and whether or not the respondent completed the ACASI interview in Spanish). In addition, we constructed three indices to measure the respondent’s engagement in risky behaviors, traditional attitudes toward sex, and attitudes toward marriage. Our modeling efforts were guided by past results and by the general hypothesis that women from traditional cultural backgrounds, featuring unfavorable attitudes toward abortion, would be more likely to underreport their abortions.

The models consistently show that older women, women with lower household income, women with a religion, and women with traditional attitudes were more likely to underreport abortions. These findings held whether we classified women based on the latent class models or simply based on their inconsistent responses across modes. In general, these findings are consistent with studies comparing reports from the National Survey of Family Growth with external “gold standard” counts. For example, one study [5] found that women who were low income, Catholic, born in the U.S., and reported their race as “Other” reported a lower proportion of their abortions.

We also fit similar models for reports about miscarriages. Although miscarriage was found as prone to underreporting as abortion, by contrast with abortion, underreporting of miscarriages was not associated with having a religion, risky substance use behaviors, or traditional sexual attitudes. It seems likely that, although miscarriages may be painful to recall and or embarrassing for many women, they are not stigmatized by traditional attitudes on sexual behavior. We also found that women with at least one child under the age of 18 in their household, women with fewer pregnancies, and women with less traditional attitudes towards marriage were more likely to underreport miscarriages as compared to other respondents, but no more likely to underreport their abortions. Race/ethnicity was the one variable related in the opposite ways to reporting on the two pregnancy outcomes. Black and Hispanic women were more likely to underreport a miscarriage than other women, but less likely to underreport an abortion (though the latter effect is not significant in every model). Perhaps miscarriage carries greater stigma in these minority communities.

The results suggest that underreporting of abortions is driven by the social stigma associated with abortion rather than, say, recall error. We did not find that women with fewer pregnancies were less likely to underreport abortions; since recall is likely to be easier for these respondents, this finding suggests that underreporting is not due to forgetting or difficulty dating an abortion event. However, an earlier study [13] found evidence suggesting that both recall and the social stigma play a role in underreporting of abortions.

We included two variables in our model in an attempt to clarify how attitudinal factors influence the dynamics of underreporting. One variable was based on the number of risky substance use behaviors (such as smoking 100 cigarettes, using cocaine, or injected drugs) each woman reported; the other was based on a series of attitudinal items assessing attitudes about sex, divorce, and gays. We thought that the two variables would be an indirect indicator of the extent to which women subscribed to the norm that abortion is wrong (which is not measured directly in the NSFG). Both variables were related to abortion reports but not with miscarriage reports, suggesting that how a woman feels about abortion is an important factor in her willingness to report one. However, it is also possible that the risky substance use behavior scale is an indicator of a willingness to report sensitive behaviors in general so that willingness to report illicit substance use may be correlated with willingness to report abortion. To accurately capture abortion, survey researchers may have to find a way to make the stigma associated with abortion less salient, thereby making it easier for people to admit having had one. Possible questionnaire design strategies include the use of random response technique [9], ballot-box technique [26], forgiving wordings [1, 27], and so on.

To evaluate the utility of using LCA to identify underreporters of abortion and miscarriage, we drew on survey reports directly to classify people into truthful reporters of abortion and underreporters of abortion. This type of classification has obvious problems since women could deny having an abortion in both CAPI and ACASI. We suspect that women denying having an abortion in both modes find questions about abortions more sensitive and stigmatic than women who reported having an abortion in the ACASI mode but not in the CAPI mode. Still, the classification results and modeling results are highly consistent with the results based on LCA model. However, LCA doesn’t seem to offer much of an advantage in exploring underreporting over the simpler method of using survey data only to identify inconsistent reporters. We suspect that, our models, like all LCAs, rely on assumptions to make estimation mathematically possible. For instance, the model in Eq 1 assumes local independence; that is, each of the indicator variables is fallible, but the errors associated with each one are uncorrelated within the latent class. Although this assumption is sometimes relaxed [20], most applications of LCA to assess errors in survey variables are based on the assumption of conditional independence [20-22]. However, in practice, respondents’ answers to one question about a topic may influence their responses to a second question measuring the same construct. They may, for example, recall their earlier answers and try to avoid appearing inconsistent. When assumptions are not met in practical survey situations, LCA estimates of error rates and prevalence are questionable [18, 28]. In the current case, the models seem to underestimate the prevalence of underreporters.

We do not see any magic bullet for markedly improving reporting about abortion or miscarriage, but do advocate including relevant attitudinal variables in future studies that attempt to measure these pregnancy outcomes. The women who underreport these outcomes are women who are likely to find them embarrassing or shameful.

Even though LCA does not seem to offer much leverage in identifying underreporters of abortion and miscarriage in the current case, it is still a useful tool and has been used for other surveys such as the Current Population Survey [21] and for other topics such as drug use [22, 23]. We encourage researchers to continue exploring and evaluating the use of LCA when there are no gold standards or error-free records available.

Question wording and factor loading for items in the traditional sex attitudes scale.

(DOCX)

Click here for additional data file.

Question wording and factor loading for items in the attitudes toward marriage scale.

(DOCX)

Click here for additional data file.

Results of logistic regression models predicting underreporting of abortion.

(DOCX)

Click here for additional data file.

Results of logistic regression models predicting underreporting of miscarriage.

(DOCX)

Click here for additional data file.

17 Feb 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Janet E Rosenbaum, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following financial disclosure:

The work reported here was supported by a grant (R01HD084473) from the Eunice Kennedy Shriver National Institute of Child Health & Human Development to The Guttmacher Institute (Laura Lindberg, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Please state what role the funders took in the study.  If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

If this statement is not correct you must amend it as needed.

Please include this amended Role of Funder statement in your cover letter; we will change the online submission form on your behalf.

3. Thank you for stating the following in the Acknowledgments Section of your manuscript:

The work reported here was supported by a grant (R01HD084473) from the Eunice Kennedy Shriver National Institute of Child Health & Human Development to The Guttmacher Institute (Laura Lindberg, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Please note that funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

The work reported here was supported by a grant (R01HD084473) from the Eunice Kennedy Shriver National Institute of Child Health & Human Development to The Guttmacher Institute (Laura Lindberg, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

4. Thank you for stating the following in your Competing Interests section:

No authors have competing interests

Please complete your Competing Interests on the online submission form to state any Competing Interests. If you have no competing interests, please state "The authors have declared that no competing interests exist.", as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now

This information should be included in your cover letter; we will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

Reviewer #3: No

Reviewer #4: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: No

Reviewer #4: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The paper seeks to contribute to our knowledge of the patterns and overall trends in underreporting of abortions and miscarriages in survey research. The authors use the National Survey of Family Growth (NSFG) to undertake their analyses. They leverage the multiple modes of data collection to investigate the characteristics of women who appear to underreport miscarriages and abortions. In addition, the authors propose using latent class analysis as a method for predicting women who are likely underreporting abortions and miscarriages. The authors use these results to compare the characteristics of women who appear to underreport miscarriages and abortions. Use of LCA and the assessment of miscarriage underreporting in the NSFG are notable contributions of the paper.

This paper has two key strengths. First, the issue of miscarriage underreporting is highly relevant and tends to be understudied due to the lack of external estimates for the “true” number of miscarriages in a population. The paper’s approach to this problem using the two forms reporting is novel and provides important insights into a topic that is difficult to study and quantify. Second, the paper’s use of LCA to identify possible under-reporters is novel and a potentially useful methodological advance for the field, especially for miscarriage, where external data are limited.

Although the topics under study are of importance and the paper makes advances in the methodological issues with measurement of these topics, the paper has two central weaknesses that limit it from fully or convincingly reaching its high potential. First, the paper is confusingly framed and situated, especially as it relates to the inclusion of the analyses related to abortion. For example, a great deal is known about the characteristics of women who underreport abortions, including specifically in the NSFG as the paper cites in the introduction. Yet, the paper motivates its analysis by suggesting little is known about the characteristics of these women. While the paper uses individual-level data, which allows for the analysis of new or additional characteristics associated with underreporting, this nuance is not clear in their introduction. More generally, the introduction and background sections of the paper are disconnected and difficult to follow (specific comments related to this can be found in the minor or stylistic issues section following these comments). Relatedly, it is unclear why the paper addresses both abortion and miscarriage underreporting. As aforementioned, abortion underreporting is well-documented, especially in the NSFG. The analysis of miscarriages, however, is quite novel. Thus, the framing of the paper is confusing and does not seem to adequately review the literature related to abortion underreporting nor make a convincing argument or present a clear rationale for why these issues are discussed in-tandem in the paper.

Second, the paper oversells the meaning of the differences between the CAPI and ACASI comparisons. That is, the paper suggests that inconsistent reporting is equivalent to underreporting, yet, there may well be something substantively different between women who never report an abortion they have had and those who sometimes will report it (on CAPI but not ACASI). There seems to be no way to know if these women are qualitatively similar or if we can draw conclusions about under-reporters as an entire group based on these inconsistent reporters. This approach is novel and interesting, but it seems to fall especially short in the case of abortion where external estimates to quantify underreporting exist. In other words, if the paper were solely focused on miscarriage, it would be more methodologically justifiable to use this approach (since a better alternative is currently lacking). Additionally, the comparison of the LCA to the inconsistent reporters is also less compelling considering these issues.

In addition to the more major issues outlined above, I also provide feedback on minor or stylistic issues by section below.

Introduction

- It is not clear why the section on LCA is presented in the introduction/background section. It seems that the materials here would be more appropriate in the analytic method section, and the limitations of LCA could be explored in the discussion section.

- The structure of the introduction could be revised for clarity as well as to more specifically discuss underreporting of abortion and miscarriages.

Methods

- It would be more accurate to label “truthful” reporter “consistent” reporters While it is unlikely someone would report an abortion or miscarriage they didn’t have, consistent reporters more accurately describes what is being observed (consistency between CAPI and ACASI)

- The choice to collapse Black and Hispanic populations together should be explicated, especially considering the import of language and religiosity in the findings. It would be stronger to separate these groups, if there is sufficient power

Results

- More descriptive table names would be helpful, especially for Table 5

- Person-first or person-centered language should be used (i.e., Black women or Black respondents or respondents of color, etc. rather than “Blacks”)

Discussion

- This section would be stronger with less reiteration of the findings and more exploration of the patterns observed and contribution

- It seems out of place to have a reference to a table in the discussion

- Inclusion of citations and references to work focused on how to improve abortion reporting or reporting of sensitive questions would be useful in the discussion

Overall, the paper addresses important issues in the study of miscarriages and abortion, especially as it relates to understanding miscarriage reporting, yet, the paper is confusingly framed and the analyses have important limitations that are not sufficiently justified or explained.

Reviewer #2: The utility of latent class analysis method, which is a powerful method for unobservable latent constructs, provides a new insight to classify respondents into “underreporter” and “reporter” groups. I think, main novelty of the study is to test utility of LCA method when estimating accurate abortion reports in a comparative way with survey reports, without any need to medical records. As you have stated in your paper, underreporting of abortions within the scope of mode differences is well-studied issue for NSFG data. However, constructing new scales on sexual attitudes and attitudes towards marriage contributes to current literature on determining factors behind underreporting behaviors of respondents.

So, I recommend revisions based on the following statements:

1- I suggest more specific manuscript title in accordance with study interests:

“Detecting Underreporters of Abortions and Miscarriages in the U.S.: A Latent Class Analysis Approach from NSFG, 2011-2015”.

2- Although the paper focuses on underreporting of abortions and miscarriages in the U.S., it is also a significant issue for other country settings. Moreover, various methodological approaches were tried to detect underreporters, and calculate accurate abortion rates at the end.

I suggest some recent works for background section. These studies include new techniques to understand underreporting for the U.S. and other countries, as well as characteristics of underreporters.

Lara, D., Strickler, J., Olavarrieta, C.D., & Ellertson, C. (2004). Measuring induced abortion in Mexico: a comparison of four methodologies. Sociological Methods & Research 32(4), 529–558.

Saraç, M., & Koç, İ. (2019). Increasing misreporting levels of induced abortion in Turkey: is this due to social desirability bias?. Journal of Biosocial Science, 52(2), 213-229.

Anderson, B.A., Katus, K., Puur, A. & Silver, B.D. (1994). The validity of survey responses on abortion: evidence from Estonia. Demography 31(1), 115–132

Tennekoon, V. S. (2017). Counting unreported abortions: A binomial-thinned zero-inflated Poisson model. Demographic Research, 36, 41-72.

Medeiros M & Diniz D (2012) Recommendations for abortion surveys using the ballot-box technique. Ciência & Saúde Coletiva 17(7), 1721–1724.

3- I suggest adding ‘abortion’, ‘miscarriage’, and ‘NSFG’ to the keywords.

4- Line 35: The authors stated that abortion is more stigmatized than miscarriage. There might be some cultural and political reasons behind that argument. The authors should provide justifications/works for that argument.

5- Line 64: ..... to understand why abortions are so badly underreported in surveys by..... Instead of that, I recommend ..... to understand factors affecting underreporting behavior of respondents by identifying the characteristics .... I think, this is more appropriate compared to “why” question, when the authors’ statistical analyses are considered.

6- Starting from Line 64: A bit difficult to follow study aims. The authors should re-organize study goals as the following. My suggestion is also more appropriate for the flow of sections about analytical strategy and results.

First goal: Evaluating utility of the LCA models in identifying underreporters of abortions/miscarriages.

Second goal: Investigating characteristics of underreporters of abortions/miscarriages.

Third goal: Determining common and differential features of respondents who underreport their abortions and miscarriages.

7- Line 124: For the readers, the authors should add a footnote for explanations of false positive and false negative errors.

8- Line 132: For a two-class LCA, three indicators are necessary for the model to be identification. The authors should add a reference here for required number of indicators to achieve model identification.

9- Line 173: … other variables. The authors should clarify on which stratification variables were used in the NSFG’s sample design, except for census division and population parameters?

10- Line 182: …. selecting one of the eligible persons for the main interview. The authors should clarify on which selection method were used to select an eligible person to interview? (Kish method, birthday method or else?)

11- Line 191: Author should give a brief information about sensitive questions in ACASI part of the questionnaire (except for abortion and miscarriage), to give a general picture for readers.

12- The authors defined “underreporter”, if a respondent who reported abortion/miscarriage in only one mode (regardless of its type; CAPI or ACASI). Instead, in accordance with the current literature, I am suggesting to define “underreporter”, if a respondent reported abortion in ACASI mode, but at the same time did not report it in CAPI mode.

As the authors discussed comprehensively in background section, respondents tend to underreport sensitive issues in face-to-face modes (i.e. CAPI) compared to self-administered modes (i.e. ACASI).

Replicating same analyses for that sub-group of respondents (ACASI reporter, CAPI underreporter) and re-organizing the paper would be useful. It is also more appropriate for the current literature and structural integrity of the paper.

13- Line 232: We used PROC LCA in SAS to fit the models.

Were the complex sample design features adjusted into LCA models?, likewise binary logistic regression analysis mentioned in Line 282. If yes, the authors should give information about that procedure, too. If no, I suggest incorporating weight, psu, and stratification variables into PROC LCA procedure in SAS.

14- Line 234: …these assignments were based on posterior class membership probabilities. This statement should be excluded. Because it has already stated within second method (Line 239).

15- Line 241: … we divided the respondent into four groups…

I suggest moving text into an organised scheme that is prepared to classify these sub-groups. It would be more useful for readers. In this scheme, the authors should provide unweighted case numbers for women groups.

16- Line 300: I guess, referring Table 2 should be replaced with Table 1 that includes survey reports.

17- Line 330: The two different methods of identifying underreporters of abortion produce very similar results. This result shows us validity/success of LCA models, authors should a sentence about that here.

18- Table 4 should include a note for representation of significance levels, as the authors added for Table 6.

19- In my opinion, results about live births should be excluded from Table 5 and scope of the paper. Abortions and miscarriages are quite relevant with literature and authors’ study objectives. There is no need to mention about live births under the results of miscarriages. Thus, I recommend restrict Table 5 only with miscarriages.

20- Line 374-377: Authors should add more discussions about the results on differential characteristics between underreporters of miscarriages and abortion. Why minority groups and marital status produce different results for two groups? Authors may refer evidence from literature here. It is also valid for discussion section (Line 428-434), too. Adding a brief explanation would be useful.

21- At the end of the discussion section, I would like to see three points:

-Can LCA method be an alternative way for other surveys in countries where there are no available medical records or, when there is only one mode of data collection?

-Could authors suggest any practical implications based on study results, especially when underreporters’ characteristics are considered? What can authors suggest for survey stages (e.g. questionnaire design, interviewer training, and data collection).

-To light future studies, which topics related to survey research are suggested by authors including use of LCA models?

22- Authors should provide exploratory factor analysis results to construct scales (like S1 Table and S2 table). The loadings of items under extracted factors and explained variances of factors should be provided in that table so that readers could follow results easily.

Reviewer #3: Broad reservation:

There is a tension in the article regarding its purpose.

The abstract and the paper look different. The abstract is focused on having two different measures of abortion and miscarriage and evaluating their consistency. This is interesting in itself and can be done focusing only on these two evaluations (and possibly analyzing with regression or means, who are the “inconsistent” reporters).

The paper is an application of latent class analysis that, in my opinion, is more debatable. Latent class analysis is about finding groups in the data, but these groups have no name. You do not get a class of “Had an abortion in last five years” vs “Did not have an abortion in last five years” as in table 2. That is pushing the method too much. In particular, when you need to introduce third unrelated variables for identification. If you bring marital status, as you do for instance, the two groups would be people who tend to have one of the marital statuses and report abortion or miscarriage in some of the dimensions and another that does not. This is different from the interpretation that I quoted. I think the paper is good while it develops the analysis in table 1, but the application of LCA is more troublesome and its interpretation should be much more cautious. I would redo or remove the LCA part.

You should think what is the purpose of the article. If you think what the abstract says is the relevant part, the article should not be as it is.

Other comments:

Based on the reservation stated above, and independently of where the article heads for, the abstract needs to be rewritten, perhaps a structured format will help. Currently, it does not talk about the results or implications of the study.

Missing recent references on underreporting in NSFG such as

Desai, S., Lindberg, L.D., Maddow-Zimet, I. et al. The Impact of Abortion Underreporting on Pregnancy Data and Related Research. Matern Child Health J 25, 1187–1192 (2021). https://doi.org/10.1007/s10995-021-03157-9

Suggestion to overcome the problems in identification:

The use of marital status as the 3rd variable makes the analysis weak since its direct association with abortion and miscarriage is unclear. Perhaps you could introduce a variable more connected to abortion and miscarriage. A possibility would be interbirth Interval (or the first birth interval for first births). The longer the last interbirth interval (or the ongoing birth interval), the more likely that missreporting is taking place. Having 3 variables measuring (somehow) the same phenomena could make the LCA interpretation more credible.

Other technical comments, although I believe the LCA and logit analysis have to be changed:

The logistic regression described in line 337 does not specify how variables have been selected. According to PLOS ONE statistical guidelines you should specify the procedure used for the identification of variables. The candidate variables should also be listed providing a rationale for the specification in the methods section. Also whether you checked for possible multicollinearity, and given the use of logistic regression, an assessment of balance. You could report it in an appendix table. It now seems that you are just putting together all those variables.

Reporting of the age-squared terms in table 4 (and 6) is wrong, showing only 0s. You should rescale so that the SE and the coefficients are shown, maybe multiplying them both by 1000. If they are all non-significant maybe it is best to remove them. Age-squared increases very much SEs given multicollinearity with age.

There is no overall statement of the probability of underreporting using weights. They could be added to tables 3 and 5.

Reviewer #4: Dear authors

I have read your manuscript with enthusiasm and enjoyed it. I have some minor comments and hope it improves the quality of the manuscript.

1. You noted that one of the aims is to understand why abortions are so badly underreported in surveys. I think it is better to say 'to quantify the level of underreporting'. I think qualitative-type works are needed to address WHY underreporting happens

2. What does term 'indicator variable' mean in LCA? Do you mean independent variables offered to the model so as to predict group membership? I guess you haven't dome that, because you did not want predictors of the group membership to be independent variables in the logistic regression models. I think it is worth to address these issues in the discussion.

3. My understanding is that LCA can provide better estimates if we include more variables in the model. Why only 2-indicator and 3-indicator variable models are developed? Wasn't is possible to include more variables in the LCA analysis?

4. LCA may classify some women to the over-report category. I liked the idea to force the model not to allocate any women to this category, and you showed that the results were robust. I am just curious how to interpret over-reported cases. Why some women should over-report abortion? Wasn't is possible to compare characteristics of truthful group with over-reported group?

5. The first paragraph in the results section can be deleted

6. How these data are calculated? "From 6.2% (n=703) to 6.6% (n=741) of the women were assigned to the “had an abortion” class, which is a bit higher than the proportion of self-reported abortions in either CAPI or ACASI modes.

7. Tables 2 and 3 show unweighted class memberships. What does unweighted mean here? Furthermore, in line 300, you noted that "weighted figures are similar". What does weighted data mean? Which weights and how applied?

8. The logistic regression outputs are not informative. I prefer to see OR, 95% CI, and P-value. Please explain on what basis variables were selected to be offered to the model? Why variable selection method, such as Backward Elimination, has not been applied? Also please report exact P-values instead of <0.10 or <0.05

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: Yes: José A. Ortega

Reviewer #4: Yes: Mohammad Reza Baneshi

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

18 Apr 2022

Responses to specific reviewer and editor comments can be found in the uploaded "Rebuttal Letter". Thanks.

Submitted filename: PLOS_Rebuttal_letter_04182022.docx

Click here for additional data file.

25 May 2022

8 Jun 2022

Dear Editor,

We have revised our manuscript to address comments from reviewers. Below we explained how we addressed each comment. We hope that this revision is satisfactory to you.

Best Regard

Ting Yan and Roger Tourangeau

Reviewers’ comments:

Reviewer #1:

-First, the comments from the authors on my second point in the first review were helpful, and the revisions begin to address this issue. However, I think it would be helpful in the discussion (around lines 435-436, where revisions were made) for the authors to either provide information or use their expertise to speculate about how similar/dissimilar the underreporters they are researching (i.e., those who report on only one CAPI or ACASI) might be to other kinds of underreporters. A sentence or two in the discussion would suffice.

-We added our speculation on differences between women who denied reporting in both modes and women who reported in ACASI but not in CAPI I in the discussion (starting with line 441).

-Second, the clarifications provided to my comment regarding “truthful reporters” was also helpful. However, in line 158, the authors state that those who report on both CAP and ACASI are “truthful” reports of abortion, while those that report in only one are likely underreporters. It would be more accurate to call the former “consistent” reporters of abortion here. Alternatively, make explicit that you are going to assume that these consistent reporters are “truthful.” A sentence making that statement would suffice. The point about the LCA classifying “truthful” reporters is acceptable, but your argumentation should be stated clearly in the paper text. I suggest adding a brief explanation as provided in the reviewer comments in the analytic strategy section.

-In Line 158, we changed the word ‘truthful’ to ‘consistent’ and added a sentence at the end of that paragraph clearly stating that we are assuming consistent reporters are truthful reporters (starting with line 166).

Other Minor issues:

Lines 45-47 [There is limited evidence… and Still…] sentences seem contradictory – maybe rephrase to clarify meaning.

-We rephrased the sentence to clarify meaning.

Line 34 – the following citation should also be included re: Add Health: Tierney, Katherine I. 2019. “Abortion Underreporting in Add Health: Findings and Implications.” Population Research and Policy Review. doi: 10.1007/s11113-019-09511-8.

-Thank you for the citation. We’ve added it.

Reviewer #2:

1- The authors stated that abortion is more stigmatized than miscarriage.

Using the findings from the study [11] to justify this argument is better. Additionally, could you add some findings from the study [12], too?

The index that Bommaraju et al. (2016) used for abortion underreporters, shows more stigma on reporting an abortion as opposed to reporting a miscarriage.

-Thank you for the suggestion. We’ve added one finding reported in Table 3 of Bommaraju et al. (2016), starting with line 40 .

2- I understood that the logistic regression models using survey data only examined characteristics of women who reported abortion/miscarriage in ACASI but not in CAPI.

But, in the text (see page 15) you said:

i. “… we fit logistic regression models to compare those who underreported their abortions (i.e. reporting group 2 above) to truthful respondents (i.e. reporting group 1).”

ii. I see group 2 above on the same page: 2) those who were assigned to the class but did not report an abortion in at least one mode.

According to the statement in ii, respondents who report their abortions/miscarriages in CAPI but not in ACASI are included in Group 2). In that case, still, the logistic models were run over women who reported abortion/miscarriage in ACASI but not in CAPI? You know, these are two different subgroups of underreporters. Could you clarify it, please? You can prefer to change your statement as follows:

“… we fit logistic regression models to compare those who underreported their abortions (reported in ACASI but not in CAPI) to truthful respondents (reported in both modes).”

- The four reporting groups described on p10 are the results of comparing women’s actual answers to the latent class they were assigned to by LCA. Both Group 1 and Group 2 were assigned to the latent class of “having an abortion.” Group 1 reported having an abortion in both modes and, as a result, they are labeled as ‘truthful reporters of abortion’. By contrast, Group 2 respondents failed to report having an abortion in either CAPI or ACASI. As a result they are labeled as ‘underreporters of abortion’.

We added some language to clarify the meaning of Group 2 (Lines 220-221, 224).

3- From your point of view based on the findings, LCA may not be regarded as a success when it is compared to the method using survey data.

However, I believe that the LCA technique could be a considerable way to detect under-reporters of abortions/miscarriages, especially for different surveys that do not provide two data collection modes to gather the same information. Could you add something like that in your discussion, considering surveys that have different designs?

I believe that journal readers would like to use the LCA technique to detect underreporters of abortions/miscarriages using data coming from different surveys.

-We’ve added a paragraph about the use of LCA for other surveys in the discussion (p24, Lines 463-467).

4- Lastly, the LCA method did not bring much advantage in identifying of underreporters of abortions in the NSFG.

However, do you suggest the use of this method when detecting underreporters of other sensitive variables in the NSFG (such as alcohol and drug use, involuntary sex and sexual disease?)

I think, the lack of advantage of the method for abortion/miscarriage underreporters may turn when underreporting of other variables are studied.

-We’ve added a couple sentence about the use of LCA for other variables in the discussion. (p24, Lines 463-467).

Submitted filename: PLOS_Rebuttal_letter_06082022.docx

Click here for additional data file.

28 Jun 2022

Detecting Underreporters of Abortions and Miscarriages in the National Study of Family Growth, 2011-2015

PONE-D-21-29890R2

Dear Dr. Yan,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Janet E Rosenbaum, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

**********

7 Jul 2022

PONE-D-21-29890R2

Detecting Underreporters of Abortions and Miscarriages in the National Study of Family Growth, 2011-2015

Dear Dr. Yan:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Janet E Rosenbaum

Academic Editor

PLOS ONE