Changes in the Perceptions of Self-weighing Across Time in a Behavioral Weight Loss Intervention.

OBJECTIVE: Changes in beliefs about self-weighing were examined across time in a behavioral weight loss intervention.
METHODS: Active duty military personnel (n = 248) enrolled in a 12-month counselor-initiated or self-paced intervention based on the Look AHEAD (Action for Health in Diabetes) Intensive Lifestyle Intervention. Using an electronic scale, participants were asked to self-weigh daily. Self-weighing perceptions were compared from baseline to 4 months (weight loss phase), from 4 months to 12 months (weight maintenance phase), and from baseline to 12 months (full intervention), as well as across time by behavioral and demographic characteristics.
RESULTS: Overall, participants perceived self-weighing as more helpful and positive, less frustrating, and making them less self-conscious after the weight loss phase. After weight maintenance, individuals believed self-weighing was less helpful and positive, more frustrating and anxiety provoking, and making them more self-conscious. However, after the intervention, participants still viewed self-weighing as more helpful and positive and less frustrating than at baseline. Weight change, self-weighing behavior prior to the intervention, and intervention condition were associated with perception change. Controlling for these influencing factors, differences in gender, BMI, age, ethnicity, and race were observed in how beliefs changed across time.
CONCLUSIONS: Results suggest engaging in a weight loss intervention promoting daily self-weighing increases positive and decreases negative beliefs about self-weighing.

RCT Entities:   Population Interventions Outcomes

Introduction

Frequent self-weighing is an effective tool for weight loss and weight gain prevention (1–7), and decreases in self-weighing are associated with subsequent weight gain (8). Extensive research refutes the common misperception that self-weighing increases long-term psychological distress and body dissatisfaction (5, 9). In fact, some studies suggest frequent self-weighing increases body satisfaction over time (10, 11).

Several studies measured perceptions of self-monitoring in adults after weight management interventions and found experiences were generally positive (3, 12–14). Steinberg and colleagues (3) measured beliefs about self-weighing in adults who completed a six-month weight loss intervention. On average, participants reported that daily self-weighing with e-scales was easy to do, easy to remember, positive, helpful, and that they would likely continue (3). Participants did not report that they found this behavior highly frustrating, anxiety provoking, or made them self-conscious. Using the same measure, consistently positive beliefs were found in African American female breast cancer survivors after a six-month weight gain prevention program (13). Additionally, Gokee Larose and colleagues (14) measured the belief that self-weighing is positive twice within a weight-gain prevention program and found scores remained positive and stable (14). However, these studies examined beliefs after engaging in an intervention (3, 12–14). Thus, potential changes in self-weighing perceptions before, during, and after a behavioral weight loss intervention are unknown.

The study purpose is to examine changes in perceptions of self-weighing during the weight loss and weight maintenance phases of a weight loss intervention in active duty military personnel using e-scales. Based on previous studies of self-weighing beliefs, in the overall sample, we hypothesize that positive perceptions will increase during weight loss (3, 12, 13). Further, based on research which suggested a stable and positive perception of self-weighing during weight gain prevention (14), we hypothesize perceptions will remain stable during weight maintenance. Given that individual experiences are embedded within a weight loss intervention, our study will examine how weight loss success, intervention condition, and self-weighing behavior prior to the intervention are associated with perception change. Additionally, due to a lack of research on demographic differences in self-weighing perceptions, our study aims to explore potential differences in gender, body mass index (BMI), age, ethnicity and race in how experiences of self-weighing differ across the intervention phases.

Methods

Participants

Active Duty Military Personnel (n=248) were recruited from Joint Base San Antonio (Table 1). Participants (≥ 18 years of age, BMI ≥ 25) had phone and email access and at least one year left in their position in San Antonio. They were required to self-monitor diet and physical activity for a week and obtain written clearance from a healthcare provider. Exclusion criteria included failing more than one military fitness test in the past 12 months, as well as having a medical condition or taking medication affecting their weight or ability to change diet or physical activity.

Table 1.

Participant Characteristics (n=248)

Demographic Characteristics	Randomized(%)
Gender
Male	122 (49%)
Female	126 (51%)
Age
< 30 years	66 (27%)
30–40 years	123 (50%)
>40 years	59 (24%)
Race
African American	49 (20%)
Caucasian	163 (66%)
Other	36 (15%)
Ethnicity
Non-Hispanic/Latino	192 (77%)
Hispanic/Latino	56 (23%)
BMI Category
Normal	1
Overweight	139 (55%)
Obese	107 (43%)

Note. Numbers represent n values

Procedure

This study was primarily approved by the Institutional Review Board (IRB) of the 59th Medical Wing in San Antonio, TX and secondarily acknowledged by the IRB at the University of Tennessee Health Science Center. After obtaining informed consent at screening, participants were randomized at the individual level using a computerized block design. Participants received either a counselor-initiated (CI) or self-paced (SP) weight-loss intervention. With both conditions, weight loss goals (10% of baseline weight), personalized calorie, and exercise goals were the same. All participants were also provided the Body Trace ™ e-scale, in which measured weights were uploaded and graphed over time into a secure personalized website. Both conditions received a 12-month manualized behavioral weight loss program based on Look AHEAD intensive lifestyle intervention. Those randomized to the CI condition received 28 individual telephone sessions with a trained interventionist over the 12-month period. Participants received interventionist feedback via email on weight, diet, and exercise self-monitoring at the same frequency as telephone sessions. Individuals in the SP condition were given a handout of available resources (i.e., individual telephone sessions and self-monitoring feedback via email when requested, lesson materials). During telephone sessions, counselors encouraged daily self-weighing as one of the most effective tools for weight loss [1-7]. Further details on the sample, study procedures, and main outcomes are described elsewhere (15, 16).

Measures

Sociodemographic characteristics.

Participants reported gender (i.e., male, female), age (i.e., <30 years, 30–40 years, >40 years), ethnicity (i.e., non-Hispanic/Latino, Hispanic/Latino), and race (i.e., Caucasian, African American, and Other) at baseline. The category “Other” included individuals who identified as American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiple races, and unknown race given infrequency of responses.

Body weight.

Weight was measured on a calibrated scale (in kilograms) at the baseline, four-month, and 12-month assessments. Height (in centimeters) was measured using a wall-mounted stadiometer. BMI was calculated using the standard formula based on baseline measurements. Participant BMI was categorized as overweight: 25.0–29.9 kg/m2 or obese: > 30 kg/m2. If clinic weight was missing for four or 12-month assessments, BodyTrace™ weights closest to targeted time points were used to maximize completeness of weight data, based on high consistency between BodyTrace™ weights and in person weight assessments (17). To define four-month weight, weight not collected in-clinic (n =36), BodyTrace™ weight that was available ± 30 days around 120 days from randomization was used (range: 90–142 days). For 12-month weight, weight not collected in-clinic (n =34), BodyTrace™ weight that was available ± 30 days around 360 days from randomization was used (range: 335–385 days). After this imputation step, missing weights at 4-months (n=13) and 12-months (n=42) assumed no change from baseline, as the most conservative estimate.

Weight change.

To define weight change (i.e., gain, loss, stable), the difference in weight from baseline to both four-months and baseline to 12-months was used. Weight change was classified: weight gain (>2.3% of baseline weight), weight loss (<2.3% of baseline weight), and weight stable (within 2.3% of baseline weight). Criterion was based on a previous study on weight maintenance (18).

Perceptions of self-weighing.

Perceptions were assessed with a questionnaire on attitudes associated with self-weighing used in previous weight loss and weight gain prevention studies (3, 14). Using an 8-point scale, participants were asked whether they found daily self-weighing to be helpful (i.e., not at all helpful (1) to very helpful (8)) and positive (i.e., not at all positive (1) to very positive (8)). Additionally, using a reverse-scored 8-point scale, participants were asked whether they found self-weighing to be frustrating (i.e., not at all frustrating (1) to very frustrating (8)), anxiety provoking (i.e., not at all anxiety provoking (1) to very anxiety provoking (8)), or made them feel self-conscious (i.e., not at all self-conscious (1) to very self-conscious (8)). Thus, a higher score for these items indicates a more negative (i.e., frustrating, anxiety provoking, making self-conscious) perception of self-weighing.

Previous self-weighing frequency.

A question from the Look AHEAD weight control practices questionnaire was administered at baseline to assess previous self-weighing behavior (i.e., “How often do you weigh yourself?”) (3, 15). Answers were never, about once a year or less, every couple of months, every month, every week, every day, and more than once per day. This study categorized participants as those who weighed themselves weekly or more frequently versus less than weekly.

Self-weighing frequency.

Self-weighing was measured continuously by measurements from the BodyTrace™ scale. Frequency was computed by adding the number of days each participant weighed during each phase of the intervention.

Analyses

Perceptions of self-weighing were compared from baseline to four-months (weight loss), four-months to 12-months (weight maintenance), and from baseline to 12-months (full intervention). This change was observed in the overall sample and compared by intervention condition, relevant weight change (i.e., four-month or 12-month outcome), and self-weighing frequency prior to the intervention using Wilcoxon Signed Rank and Kruskal-Wallis tests. To explore demographic differences, linear regression models examined change by gender, BMI category, age, ethnicity and race controlling for these influencing covariates (i.e., weight change, intervention condition, previous self-weighing). In additional analyses, models examined demographic differences controlling for intervention condition, weight change, and frequency of e-scale self-weighing during relevant intervention phase (Table S1). To observe longitudinal profiles across the full intervention, random coefficient models predicted change over time controlling for baseline covariates (i.e., intervention, previous self-weighing frequency). Finally, a Spearman correlation coefficient was computed to assess the relationship between change in each belief and frequency of self-weighing during each intervention phase.

Results

Randomized participants were diverse in gender, BMI category, race, ethnicity and age (Table 1). Frequency of self-weighing behavior across behavioral and demographic characteristics are provided (Table 2).

Table 2.

Frequency of Self-Weighing

	Weight Loss Phase		Weight Maintenance Phase		Full Intervention
	N	M (SD)	N	M (SD)	N	M (SD)
Overall SampleGender	246	71.95 (34.47)	246	78.03 (66.96)	246	149.98 (94.31)
Male	121	72.36 (36.52)	121	77.55 (67.02)	121	149.92 (95.64)
FemaleAge	132	71.54 (32.52)	125	78.49 (67.17)	125	150.03 (93.39)
<30 years	66	62.73 (35.34)	66	62.17 (62.53)	66	124.89 (90.24)
30–40 years	122	74.29 (34.03)	122	80.11 (66.52)	122	154.40 (92.99)
>40 yearsRace	58	77.52 (32.92)	58	91.69 (70.20)	58	169.21 (97.18)
Caucasian	161	70.01 (34.41)	161	77.20 (64.71)	161	147.22 (92.20)
African American	49	77.61 (31.25)	49	77.90 (65.08)	49	155.51 (88.35)
OtherEthnicity	36	72.89 (38.80)	36	81.89 (80.10)	36	154.78 (112.22)
Non-Hispanic/Latino	191	73.27 (34.84)	191	81.1 (65.65)	191	154.37 (93.57)
Hispanic-LatinoBMI Category	55	67.36 (33.08)	55	67.35 (70.89)	55	134.71 (96.16)
Overweight	114	72.24 (33.76)	114	78.30 (69.87)	114	150.54 (95.95)
ObeseWeight Change	132	71.70 (35.21)	132	77.80 (64.61)	132	149.49 (93.24)
Weight Loss	90	85.71 (29.43)	90	113.11 (67.43)	90	198.82 (90.60)
Weight Stable	100	66.47 (34.98)	100	57.12 (59.79)	100	123.59 (85.66)
Weight GainIntervention	56	59.61 (34.21)	56	58.98 (54.78)	56	118.59 (83.42)
Counselor-Initiated	114	88.15 (27.97)	114	97.11 (68.13)	114	185.27 (89.31)
Self-PacedPrevious Self-Weighing	132	55.74 (32.78)	132	58.94 (60.23)	132	114.68 (85.88)
At least weekly	143	73.67 (33.3)	143	82.2 (69.92)	143	155.87 (96.1)
Less than weekly	102	69.25 (36.12)	102	72.15 (62.79)	102	141.4 (91.98)

Note. Mean (M); Standard deviation (SD)

Change in Self-Weighing Perceptions in the Overall Sample

Weight loss phase (0–4 months).

Participants reported self-weighing as more helpful [Mean (M) = 1.56, p<.0001] and more positive (M=1.66, p<.0001), as well as less frustrating (M= −1.02, p<.0001) and making them less self-conscious (M= −0.76, p<.0001) after the weight loss phase (Table 3). Belief that self-weighing was anxiety-provoking did not significantly change (baseline score M=3.33).

Table 3.

Description of Self-Weighing Perception Scores Across Intervention Phases by Behavioral Characteristics

Perception ChangeM (SD)	Overall Sample		Weight Change		Intervention Group		Previous Self-Weighing
Helpful		Loss	Stable	Gain	Counselor	Self-Paced	<Weekly	≥Weekly
Baseline	5.15 (1.93)	4.82 (1.93)	5.32 (1.94)	5.31 (1.88)	4.91 (2.04)	5.39 (1.79)	4.68 (2.01)**	5.49 (1.81)**
Weight Loss	1.56 (2.09)***	1.94 (2.05)** †	1.17 (2.09)** †	1.13 (2.00)** †	2.07 (2.05) ***	.93 (1.98) ***	1.97 (2.02)*	1.28 (2.10)*
Weight Maintenance	−.28 (1.52)*	.15 (1.19)**	−.72 (1.60)**	−.55 (1.84)**	−.26 (1.47)	−.30 (1.61)	0.00 (1.30)*	−.50 (1.66)*
Full Intervention	1.31 (2.27)***	2.15 (1.94)***	.49 (1.97)***	.85 (2.67)***	1.85 (2.17)***	.62 (2.21)***	2.03 (2.24)***	.81 (2.16)***
Positive
Baseline	4.12 (1.98)	3.77 (1.96)	4.23 (2.00)	4.44 (1.92)	4.00 (2.06)	4.24 (1.89)	3.80 (1.99)*	4.35 (1.94)*
Weight Loss	1.66 (2.20)***	2.44 (1.84)*** †	1.03 (2.15)*** †	−.13 (2.53)*** †	2.17 (2.12)***	1.03 (2.13)***	2.23 (2.15)**	1.27 (2.15)**
Weight Maintenance	−.38 (1.47)**	−.12 (1.38)	−.66 (1.37)	−.55 (1.76)	−.38 (1.30)	−.39 (1.72)	−.33 (1.49)	−.41 (1.47)
Full Intervention	1.18 (2.13)***	2.01 (2.02)***	.92 (2.02)***	.03 (1.86)***	1.78 (2.26)***	.41 (1.68)***	1.69 (2.30)*	.83 (1.94)*
Frustrating
Baseline	4.62 (2.12)	4.51 (2.10)	4.59 (2.05)	4.58 (2.13)	4.55 (2.10)	4.68 (2.14)	4.63 (2.18)	4.60 (2.08)
Weight Loss	−1.02 (2.26)***	−1.38 (2.26) †	−.68 (2.07) †	−.44 (2.87) †	−1.11 (2.43)	−.91 (2.05)	−1.04 (2.03)	−1.01 (2.42)
Weight Maintenance	.36 (2.01)*	−.21 (1.67)**	.83 (1.91)**	.93 (2.55)**	.32 (1.97)	.42 (2.09)	.11 (1.84)	.56 (2.14)
Full Intervention	−.61 (2.25)**	−1.36 (2.02)***	−.35 (2.09)***	.40 (2.39)***	−.83 (2.45)	−.34 (1.93)	−.82 (2.19)	−.47 (2.28)
Anxiety Provoking
Baseline	3.33 (2.10)	3.47 (2.04)	3.22 (2.15)	3.33 (2.08)	3.32 (2.09)	3.34 (2.11)	3.41 (2.20)	3.27 (2.03)
Weight Loss	−.22 (1.99)	−.43 (1.90) †	−.04 (1.96) †	.19 (2.61) †	−.30 (1.92)	−.13 (2.09)	−.24 (1.94)	−.21 (2.04)
Weight Maintenance	.31 (1.99)	−.16 (1.91) *	.89 (1.97) *	.48 (1.99)*	.21 (1.84)	.47 (2.20)	.06 (2.16)	.50 (1.84)
Full Intervention	−.06 (2.28)	−.74 (2.06)**	.25 (2.46)**	.75 (2.13)**	−.18 (2.22)	.10 (2.36)	−.34 (2.46)	.14 (2.14)
Self-Conscious
Baseline	4.58 (2.36)	4.71 (2.19)	4.58 (2.51)	4.37 (2.31)	4.62 (2.41)	4.54 (2.32)	4.72 (2.31)	4.48 (2.40)
Weight Loss	−.76 (2.22)***	−1.20 (2.25)** †	−.39 (1.87)** †	.25 (3.02)** †	−.90 (2.24)	−.58 (2.20)	−.84 (2.10)	−.70 (2.31)
Weight Maintenance	.62 (1.80)***	.26 (1.57)	.77 (2.06)	1.21 (1.72)	.53 (1.93)	.75 (1.58)	.37 (1.84)	.85 (1.73)
Full Intervention	−.15 (2.24)	−.72 (1.76)***	−.20 (2.65)***	0.95 (2.06)***	−.51 (2.30)**	.32 (2.07)**	−.43 (2.22)	.05 (2.24)

Note.

p<.001,

p<.01,

p<.05;

Mean (M); Standard deviation (SD); Wilcoxon Signed Rank and Kruskal-Wallis tests measured mean change differences overall and between behavioral factors;

Weight change based on 4-month outcome

Weight maintenance phase (4–12 months).

Compared to scores after the weight loss phase, participants reported self-weighing as less helpful and less positive after weight maintenance (p=.036, p=.0024, respectively) (Table 3). Further, they found that self-weighing was more frustrating (p=.035) and making them more self-conscious (p<.0001). Belief that self-weighing was anxiety-provoking remained unchanged.

Full intervention (0–12 months).

Participants believed self-weighing was more helpful (M=1.31, p<.0001) and more positive (M=1.18, p<.0001), as well as less frustrating (M= −0.61, p=.0004) after the full intervention (Table 3). Beliefs that self-weighing was anxiety provoking and makes you self-conscious remained unchanged.

Perceptions by Behavioral Characteristics

At baseline, there were no differences in self-weighing perceptions by weight change outcomes at four-months (data not shown) or 12-months (Table 3).

Across the weight loss phase, those who lost weight (i.e., at four-months) viewed self-weighing as more helpful (p=.013), more positive (p<.00001), and making them less self-conscious (p=.008) compared to those who gained or remained weight stable (Table 3). Across weight maintenance, individuals who lost weight (i.e., at 12-months) perceived self-weighing as more helpful (p<.0001), more positive (p<.0001), less frustrating (p=.0004), less anxiety-provoking (p=.007) and making them less self-conscious (p=.0006) than those who gained or remained weight stable. Over the full intervention, those who lost weight (i.e., 12-months) developed views that self-weighing was more helpful (p<.0001) and positive (p<.0001), as well as less frustrating (p=.0004), anxiety-provoking (p=.007), and making them less self-conscious (p=.0006). No other differences were observed.

Intervention group.

Participants in both interventions had similar baseline perceptions (Table 3). Across the weight loss phase, the CI group believed self-weighing was more helpful (p=.0001) and more positive (p=.0004) than the SP group (Table 3). Over the full intervention, the CI group perceived self-weighing as more helpful (p=.0003) and positive (p<.0001), and making them less self-conscious (p=.004) than the SP group. No other differences in beliefs were observed.

Previous self-weighing.

At baseline, individuals who weighed weekly or more frequently prior to the intervention indicated that self-weighing was more helpful (p=.002) and positive (p=.03) than those who weighed less often (Table 3). Across weight loss phase, those who weighed less than weekly prior to the intervention believed self-weighing was more helpful (p=.046) and more positive (p=.004) than those who weighed more frequently (Table 3). Across weight maintenance, those who weighed less than weekly prior to the intervention believed self-weighing was more helpful (p=.014) than those who weighed more frequently (Table 3). Across the full intervention, those who weighed less than weekly prior to the intervention developed views of self-weighing as more helpful (p=.0003) and positive (p=.033) than those who weighed more frequently.

Perceptions by Demographic Characteristics

BMI.

At baseline, participants with overweight believed self-weighing was more positive (p=.027) and less frustrating (p=.029) than participants with obesity (Table 4). Controlling for covariates, across the weight loss phase, individuals with obesity believed self-weighing was more helpful (p=.027) and more positive (p=.006) than those with overweight (Table 4). Across weight maintenance, however, those with obesity believed self-weighing was less positive (p=.02) and more anxiety provoking (p=.022). There were no BMI differences for other beliefs at baseline or across time (Table 4).

Table 4.

Description of Self-Weighing Perception Scores Across Intervention Phases by Demographic Characteristics

Perception ChangeM (SD)	BMI		Gender			Age		Ethnicity			Race
Helpful	≥ 30	> 25	Male	Female	<30 years	30–40 years	>40 years	Non-	Hispanic/Latino	African American	Caucasian	Other
Baseline	5.00 (1.93)	5.32 (1.93)	5.20 (1.93)	5.10 (1.94)	4.97 (1.86)	5.17 (2.01)	5.31 (1.86)	Hispanic/Latino 5.23 (1.92)	4.89 (1.96)	5.08 (1.96)	5.02 (1.87)	5.81 (2.07)
Weight Loss	1.84 (.24)* †	1.20 (.25)* †	1.56 (.24) †	1.53 (.25) †	1.49 (.33) †	1.46 (.25) †	1.76 (.32) †	1.49 (.22) †	1.71 (.34) †	1.79 (.36) †	1.62 (.23) †	.87 (.39) †
Maintenance	−.38 (. 18)†	−.26 (.19) †	−.31 (.18) †	−.35 (.19) †	−.39 (.28) †	−.26 (.18) †	−.41 (.25) †	−.25 (.15) †	−.63 (.28) †	−.20 (.30)* †	−.17 (.16)* †	−1.08 (.30)* †
Full Intervention	1.44 (.23) †	1.00 (.25) †	1.02 (.23) †	1.48 (.24) †	1.04 (.36) †	1.39 (.24) †	1.13 (.33) †	1.20 (.19) †	1.39 (.36) †	1.54 (.38)* †	1.39 (.20)* †	.27 (.40)* †
Posituve
Baseline	3.87 (1.96)*	4.41 (1.97)*	4.52 (1.80) ***	3.74 (2.07) ***	3.97 (1.95)	4.29 (1.99)	3.95 (2.00)	4.14 (1.95)	4.07 (2.09)	4.80 (2.04)***	3.71 (1.79)***	5.06 (2.18)***
Weight Loss	1.61 (.24)** †	.81 (.25)** †	1.08 (.24) †	1.41 (.25) †	1.33 (.32) †	.93 (.25) †	1.74 (.32) †	1.23 (.22) †	1.25 (.34) †	1.15 (.35)** †	1.49 (.22)** †	.26 (.38)** †
Maintenance	−.69 (.17)* †	−.11 (.19)* †	−.48 (.18) †	−.37 (.19) †	−.38 (.28) †	−.43 (.18) †	−.46 (.25) †	−.41 (.15) †	−.51 (.29) †	−.39 (.31) †	−.50 (.16) †	−.20 (.31) †
Full Intervention	1.11 (.21) †	.91 (.23) †	.70 (.21)* †	1.35 (.22)* †	1.13 (.33) †	.71 (.22) †	1.53 (.30) †	1.07 (.18) †	.82 (.34) †	.44 (.36)* †	1.28 (.19)* †	.54 (.38)* †
Frustrating
Baseline	4.89 (2.17)*	4.30 (2.02)*	4.20 (2.03)**	5.02 (2.13)**	5.00 (2.03)	4.43 (2.09)	4.58 (2.25)	4.69 (2.10)	4.36 (2.18)	4.55 (2.35)**	4.85 (1.99)**	3.64 (2.13)**
Weight Loss	−.93 (.27) †	−.69 (.28) †	−.81 (.27) †	−.82 (.28) †	−1.09 (.37) †	−.67 (.28) †	−0.82 (.36) †	−.89 (.24) †	−.59 (.38) †	−.83 (.40) †	−.91 (.26) †	−.38 (.44) †
Maintenance	.54 (.23) †	.40 (.26) †	.41 (.24) †	.55 (.25) †	.55 (.38) †	.50 (.24) †	.38 (.33) †	.46 (.20) †	.54 (.38) †	−.11 (.41) †	.63 (.21) †	.43 (.41) †
Full Intervention	−.48 (.24) †	−.42 (.26) †	−.44 (.24) †	−.46 (.25) †	−.55 (.37) †	−.32 (.25) †	−.61 (.34) †	−.60 (.20) †	.10 (.37) †	−.83 (.41) †	−.44 (.21) †	−.04 (.43) †
Anxiety-Provoking
Baseline	3.46 (2.12)	3.18 (2.07)	2.94 (1.97)**	3.70 (2.16)**	3.41 (2.04)	3.11 (2.07)	3.68 (2.19)	3.39 (2.10)	3.11 (2.08)	3.18 (2.31) *	3.55 (2.05) *	2.53 (1.84) *
Weight Loss	−.17 (.24) †	−.01 (.25) †	−.11 (.24) †	−.07 (.25) †	−.56 (.33) †	.26 (.25) †	−.29 (.32) †	−.24 (.22) †	.39 (.34) †	.21 (.36) †	−.26 (.23) †	.26 (.40) †
Maintenance	.72 (.23)* †	−0.04 (.25)* †	.44 (.24) †	.30 (.25) †	.02 (.37) †	.61 (.24) †	.18 (.33) †	.29 (.20) †	.67 (.38) †	.48 (.41) †	.51 (.21) †	−.26 (.40) †
Full Intervention	.26 (.24) †	−.19 (.27) †	.06 (.25) †	.05 (.26) †	−.38 (.38) †	.46 (.25) †	−.36 (.34) †	−.18 (.20)* †	.90 (.38)* †	−.03 (.42) †	.10 (.22) †	−.03 (.44) t
Self-Conscious
Baseline	4.71 (2.38)	4.42 (2.34)	3.93 (2.18)***	5.21 (2.37)***	4.82 (2.35)	4.38 (2.36)	4.73 (2.37)	4.53 (2.40)	4.75 (2.24)	4.37 (2.62)	4.72 (2.26)	4.25 (2.30)
Weight Loss	−.47 (.26) †	−.41 (.28) †	−.28 (.27) †	−.62 (.27) †	−.55 (.36) †	−.40 (.27) †	−.41 (.36) †	−.52 (.24) †	−.20 (.37) †	−.51 (.39) †	−.46 (.25) †	−.26 (.43) †
Maintenance	.83 (.21) †	.59 (.23) †	.81 (.22) †	.63 (.22) †	.54 (.33) †	1.05 (.21) †	.22 (.29) †	.65 (.18) †	.97 (.34) †	.85 (.37) †	.77 (.19) †	.41 (.37) †
Full Intervention	.04 (.24) †	−.06 (.26) †	.20 (.24) †	−.22 (.25) †	−.12 (.37) †	.26 (.24) †	−.42 (.34) †	−.05 (.20) †	.16 (.38) †	.40 (.40) †	−.01 (.21) †	−.42 (.43) †

Note.

p<.001,

p<.01,

p<.05;

Mean (M); Standard deviation (SD); Wilcoxon Signed Rank and Kruskal-Wallis tests measured mean change differences at baseline between demographic factors; Linear regression models observed demographic differences in change across each intervention phase controlling for covariates;

Means adjusted for intervention condition, baseline self-weighing frequency, and weight change at the end of the relevant intervention phase (i.e., 4-month outcome, 12-month outcome)

Gender.

At baseline, men reported self-weighing was more positive (p=.001), less frustrating (p=.002), less anxiety provoking (p=.005), and made them less self-conscious (p<.0001) than women (Table 4). There were no other differences at baseline or across weight loss or maintenance phases (Table 4). Adjusting for covariates, across the full intervention, women developed the view of self-weighing as more positive (p=.03). No other differences were found across the full intervention.

Age.

Baseline scores showed similar beliefs of self-weighing across age groups, and no differences were observed for beliefs over time (Table 4).

Ethnicity.

At baseline, scores showed similar perceptions regardless of ethnicity (Table 4). Adjusting for covariates, individuals who identified as Hispanic/Latino believed self-weighing was more anxiety-provoking (p=.013) than those non-Hispanic/Latino after the full intervention (Table 4). No other differences were found.

Race.

At baseline, participants who identified as Caucasian reported self-weighing was less positive (p<.0001) compared to other racial identities (Table 4). Further, participants classified as Other race perceived self-weighing as less frustrating (p=.009) and less anxiety provoking (p=.017) at baseline than other racial identities. There were no other baseline differences. Adjusting for covariates, those classified as Other race believed self-weighing was less helpful across weight maintenance (p=.03) and the full intervention (p=.022) compared to those with other racial identities. Across the full intervention, those identified as Caucasian developed self-weighing beliefs of being more positive (p=.044) than those with other racial identities.

Longitudinal Profiles

In random coefficient models, interactions of time and behavioral characteristics were consistent with linear regression models (Figure 1). To differentiate between intervention and weight loss success, models explored effects of intervention group and time, controlling for weight change (i.e., 12-months), for all beliefs. Interactions of time and intervention condition remained for beliefs that self-weighing is helpful (p=.0041) and positive (p=.0054), such that those in the CI group exhibited a steeper increase in the beliefs that self-weighing is helpful and positive compared to the SP group over the full intervention. No significant interactions were found for other beliefs.

Figure 1.

Modeling Change in Self-Weighing Perceptions across the Intervention

Random Coefficient models observed change in self-weighing perceptions by behavioral and demographic factors. Self-weighing perceptions (x-axis) in order: Helpful, Positive, Frustrating, Anxiety-Provoking, Makes You Self-Conscious

Factors of interest (y-axis) in order: overall sample, previous self-weighing frequency (i.e., < weekly, ≥ weekly), weight change (i.e., gain, loss, stable), intervention condition (i.e., counselor-initiated, self-paced), BMI category (i.e., overweight, obese), gender, age (i.e., <30, 30–40, >40), ethnicity (i.e., non-Hispanic/Latino, Hispanic/Latino), race (i.e., Caucasian, African American, Other)

Models of demographic differences (i.e., BMI, gender, age, ethnicity, and race) adjusted for baseline covariates (i.e., intervention condition, previous self-weighing frequency)

A higher score indicates a more positive (i.e., helpful, positive) perception of self-weighing. Scores for negative beliefs were reverse scored, such that, a higher score indicates a more negative (i.e., frustrating, anxiety provoking, making self-conscious) perception of self-weighing.

Adjusting for baseline covariates, coefficient models predicted profile differences by BMI, gender, ethnicity, and race similar to differences found in linear regression models (Figure 1). However, inconsistent with null findings in regression models, time and age interacted in the belief that self-weighing is anxiety provoking (p =.027) (Figure 1). Specifically, those between the ages of 30 and 40 years of age exhibited a steeper increase in the belief that self-weighing is anxiety provoking compared to other ages over the full intervention.

Self-Weighing Perceptions and Self-Weighing Behavior

Weight loss phase.

There was a positive relationship between increased belief that self-weighing is helpful and number of days participants self-weighed during the weight loss phase (r =0.35, p<.0001) (Table 5), such that developing stronger beliefs that self-weighing is helpful correlated with more frequent self-weighing. There was a negative relationship between increased belief that self-weighing is positive and days weighed (r = −0.19, p=.007), such that developing stronger beliefs that self-weighing is positive correlated with less frequent self-weighing. Further, there was a negative relationship between increased belief that self-weighing is frustrating and days participants weighed (r = −.15, p=.020), such that developing stronger views that self-weighing is frustrating correlated with less frequent weighing. There were no correlations between change in other beliefs and self-weighing frequency.

Table 5.

Association of Change in Self-Weighing Perceptions and Self-Weighing Frequency

	Self-Weighing Frequency
	Weight Loss	Weight Maintenance	Full Intervention
Weight Loss
Helpful	35***	32***	.35***
Positive	−.19**	−.19**	−.19**
Frustrating	−.15*	−.20**	−.19**
Anxiety-Provoking	−.08	−.09	−.09
Self-Conscious	.08	.06	.07
Weight Maintenance
Helpful	−.01	.02	−.01
Positive	.00	.00	−.00
Frustrating	−.02	−.12	−.09
Anxiety-Provoking	−.06	−.08	−.07
Self-Conscious	−.07	−.05	−.06
Intervention
Helpful	.29***	.26**	.29***
Positive	−.20*	−.20*	−.22**
Frustrating	−.13	−.15	−.16
Anxiety-Provoking	−.16	−.15	−.16
Self-Conscious	.09	.06	.07

Note.

p<.001,

p<.01,

p<.05;

Spearman correlation coefficients measured associations; Numbers represent R values

Weight maintenance phase.

There were no associations between change in perceptions and self-weighing frequency during weight maintenance (Table 5).

Full intervention.

Across both phases, there was a positive relationship between increased belief that self-weighing is helpful and days participants self-weighed (r= 0.29, p=.0005), such that developing stronger beliefs that self-weighing is helpful correlated with more frequent self-weighing (Table 5). There was a negative relationship between increased belief that self-weighing is positive and days participants self-weighed (r = −0.22, p=.0086), such that developing stronger beliefs that self-weighing is positive correlated with less frequent self-weighing. There were no associations between change in other beliefs and self-weighing frequency.

Discussion

Current findings are the first, to our knowledge, to track the progress of beliefs specifically about self-weighing behavior across different phases of a weight loss intervention. Consistent with our hypothesis, overall, positive perceptions of self-weighing increased and negative perceptions decreased during the weight loss phase. However, positive perceptions decreased and negative perceptions increased both slightly, albeit significantly, during the weight maintenance phase. Yet, after intervention completion, active duty personnel still perceived self-weighing as more positive and less negative than they did at baseline. Results indicate that when observing the sample as a whole, participation in a weight loss intervention, especially the weight loss phase, increased positive beliefs and decreased negative beliefs about weight self-monitoring behavior, although this effect was stronger in the more intensive version of the intervention. Consistent with previous qualitative findings, results suggest that participants find that self-weighing is not as frustrating as expected and that it can be helpful and positive in achieving goals after engaging in an intervention promoting daily self-weighing (19).

Ratings after the weight loss phase (i.e., four-months) were similar to a previous study which measured these perceptions at six-months; specifically, overall mean scores in the beliefs that self-weighing is helpful, positive, anxiety provoking, and makes you self-conscious, were similar to those reported by Steinberg and colleagues (3). However, the slight decrease over the maintenance period in the belief that self-weighing is positive was inconsistent with previous findings (14), in which this perception remained stable in a weight gain prevention program. Further, compared to African American breast cancer survivors in a weight management program (13), this study sample’s perceptions about self-weighing were less positive and more negative. Perhaps, current ratings were specific to those experienced in an intervention that included both weight loss and weight maintenance phases, as well as within a primarily (66%) Caucasian sample.

Given the potential for different experiences of self-weighing throughout the intervention, our study aimed to disentangle behavioral characteristics associated with change in beliefs. Weight change, self-weighing behavior prior to the intervention, and the intensity of the intervention were associated with how self-weighing perceptions changed. Specifically, those who weighed less frequently at baseline perceived self-weighing as more helpful and positive over time than those who previously weighed more often. Participants who weighed infrequently prior to the intervention might have been unfamiliar with self-weighing as a weight loss tool, and perhaps, these participants were more positively influenced by exposure to this behavior. Not surprisingly, those who lost weight perceived self-weighing as more positive and less negative across the intervention. Additionally, those in the more intensive CI intervention group believed self-weighing was more positive and helpful than the SP group, regardless of weight loss success. Counselor encouragement and feedback about self-weighing might have facilitated a more positive experience when engaging in this behavior even within individuals who were unsuccessful with weight loss.

Adjusting for all covariates, no age differences were found in beliefs across intervention phases. Yet, longitudinal profiles, adjusting for baseline covariates, suggested that those between the ages of 30 and 40 experienced self-weighing as more anxiety provoking compared to other ages. Findings might be unique to the military population, in which individuals between the ages of 30 and 40 are closer to retirement than in other professions (20). Failure to pass annual fitness tests increases an individual’s risk of being discharged from their job and losing medical and pension benefits (21). Thus, perhaps these active duty personnel between 30 and 40 years experienced increased anxiety with self-weighing as the weight loss intervention neared completion in anticipation of a fitness test.

Participants with obesity, despite perceiving self-weighing as more helpful and positive during the weight loss phase, experienced this behavior as less positive and more anxiety provoking over weight maintenance compared to those with overweight. These findings indicate that participants with obesity might benefit from increased support in regard to self-weighing during weight maintenance periods. This BMI difference might explain why the mean rating of self-weighing as frustrating in the present study (i.e., M =3.48) was higher compared to the mean reported by Steinberg and colleagues (3) (i.e., M=2.4). Steinberg and colleagues (3) measured perceptions in only adults with overweight. Perhaps, those with obesity, although believing self-weighing is helpful, might perceive more negative experiences with this behavior. Further, women developed a more positive view of self-weighing across the intervention compared to men. Women initially believed self-weighing was less positive and more negative prior to the intervention, thus, they experienced a greater change in their perception of this behavior over time.

Despite similar beliefs at baseline, those identified as Hispanic/Latino perceived self-weighing as more anxiety provoking compared to those non-Hispanic/Latino. Additionally, individuals classified as Other race believed self-weighing was less helpful across intervention phases compared to other racial identities. Although participants identified as Caucasian reported less positive beliefs prior to the intervention, they were more likely to report a positive experience of self-weighing over time compared to other racial identities. Findings suggest adults identified as ethnic or racial minorities experience self-weighing more negatively across a behavioral trial. Future interventions might provide additional support and resources, specifically in regard to self-weighing, for individuals identifying as racial or ethnic minorities. However, due to the combined different racial identities, interpretation of results related to the Other race is difficult. Future research should utilize larger samples of individuals who identify as American Indian/Alaska Native, Asian, Native Hawaiian/Pacific Islander, multiple races, and unknown race in order to draw conclusions.

With regard to the correlations between perception change and actual self-weighing behavior, perhaps, participants who weighed more frequently during the intervention were more likely to perceive self-weighing as helpful, yet also felt that this behavior was less positive and more frustrating. Interestingly, other perceptions were not associated with self-weighing behavior. Although these correlations were weak, results suggest that an increased belief that self-weighing is helpful might facilitate more frequent self-weighing behaviors more so than other beliefs.

The current study was conducted in a military population which might limit generalizability to the civilian sector. Personnel are generally younger and have access to additional resources to assist in weight loss that are not accessible to the general public (e.g., free fitness centers, healthcare). Further, results might not be representative of adults with severe health and mental health conditions who are excluded from enlistment. In the military, fitness is perceived as a critical factor for employment and is assessed regularly in annual evaluations. Despite these unique characteristics of the military, the prevalence of overweight and obesity in the U.S. military is similar to the civilian population. Approximately 60% of active duty personnel are affected by overweight and obesity, which is a critical concern for the U.S. military (22). This diverse population of military personnel provided a unique opportunity to examine race differences, as well as gender differences, dissimilarly from previous studies which examined primarily female samples (3, 13–14, 19).

Missing in-clinic weight for some participants was a limitation. Although minimized by utilizing BodyTrace™ weight, there were cases in which weights were not available. The current study used a criterion of 2.3% to define weight change based on previous research (18). However, other studies suggest alternative definitions (e.g., 3%) (23). Importantly, although the protocol clearly encouraged daily weighing, information on how frequently self-weighing was discussed during the telephone sessions was not collected. In addition, those in the CI condition more frequently received encouragement to self-weigh. Finally, although the questionnaire of self-weighing perceptions has been used in previous research, it was not previously validated. Future research should assess reliability and validity of this measure.

Conclusions

Despite the weight loss benefits of self-monitoring, there is often a reluctance to engage in self-weighing behaviors. Current results suggest completing a weight loss intervention, which promoted daily self-weighing, overall increased positive perceptions and decreased negative perceptions of self-weighing. Importantly, weight loss success, prior lack of experience with self-weighing, and increased intervention intensity were associated with a more positive experience of self-weighing over time. Controlling for these influencing factors, differences in BMI, gender, age, ethnicity, and race were also observed in how these experiences changed across the intervention.