PSYCH 664 Week 3 Sampling and Measurement Example

Who Gets Invited and How Loneliness Gets Counted: Sampling, Power and Measurement for the Senior Phone-Call Study

[Student Name]

University of Phoenix

PSYCH/664: Research Methods and Statistics in Psychology

Week 3 Assignment

[Instructor Name]

[Date]

The senior community, its residents and the study plan are composites written for a model paper; measures and methodological sources come from the references listed.

A strong design can still fail if the wrong people are studied or outcomes are measured poorly. This paper plans sampling and measurement for the study of weekly phone calls from college students to lonely residents of our senior community.

Population and Sampling Frame

The target population is older adults living in senior housing who experience loneliness. The accessible population is residents of our community in Sun City, Arizona, about 320 people in independent and assisted living. The sampling frame is the resident directory. Because the study tests an intervention within one community, I will invite all residents who meet eligibility criteria rather than drawing a random sample; this is a census of eligible residents, followed by random assignment to groups. Generalizing beyond our community will rely on describing the sample carefully rather than on probability sampling.

Eligibility

Inclusion criteria: age seventy or older; able to speak and hear well enough for phone conversation, with amplified phones available; able to give informed consent, assessed with a brief consent comprehension check; and a score indicating at least moderate loneliness at screening. Exclusion criteria: residents in hospice care or with severe cognitive impairment that prevents consent. Including residents with mild hearing loss through amplified phones avoids excluding a group at high risk for loneliness.

How Many Participants?

Faul et al. (2007) described G*Power 3, a free program for statistical power analysis that supports many tests, including t tests, F tests and chi-square tests. Users specify the test, the expected effect size, alpha and desired power, and the program computes the required sample size, or the power for a given sample. The authors emphasized choosing effect sizes based on prior research or the smallest effect of practical importance.

Prior phone-call trials suggest moderate effects on loneliness. Using an expected standardized difference of 0.65 between groups at twelve weeks, a two-tailed significance level of .05 and an 80 percent chance of detecting the effect, a two-group comparison requires about thirty-nine residents per group. Allowing for fifteen percent attrition, common in older adult studies, I will aim to enroll about ninety residents, forty-five per group. If fewer eligible residents enroll, the study may detect only larger effects, which I will report honestly.

With ninety residents, the study can find a moderate effect; with forty, it could miss a real one.

Who Might Be Missed

Even a census of eligible residents can miss people. Residents with moderate dementia cannot consent and are excluded, though some experience loneliness. Residents who distrust surveys or dislike talking on the phone may decline, and their absence would make the sample less lonely and more sociable than the community as a whole. Men are a minority in our community and may be less likely to volunteer for a program framed around loneliness. I will record the number invited, screened, eligible and enrolled, along with basic characteristics of those who decline when they are willing to share them, so readers can see how representative the sample is.

Effect Size Assumptions

The choice of 0.65 as the expected effect deserves scrutiny. Published trials of brief phone programs reported moderate effects, but small, early trials tend to overestimate effects, and our residents already have on-site activities, which may leave less room for improvement. If the true effect is closer to 0.4, about one hundred residents per group would be needed, more than our community can provide. I will therefore treat the study as a well-designed pilot whose effect estimate, with its confidence interval, can guide a larger multisite trial.

Recruitment

Recruitment will include a presentation at the monthly residents' meeting, flyers in large print, notes in mailboxes and personal invitations from staff who know residents who rarely attend activities. The last strategy matters because the loneliest residents may be least likely to respond to flyers.

Measuring Loneliness

Russell (1996) reported the development and psychometric properties of the third version of the UCLA Loneliness Scale, a twenty-item measure with simplified wording and balanced positively and negatively worded items. Across samples of college students, nurses, teachers and elderly adults, internal consistency was high, with coefficients from .89 to .94, and test-retest reliability over one year was .73. Scores correlated with measures of social support, depression and health in expected directions and showed discriminant validity from measures of other constructs.

Hughes et al. (2004) developed a three-item loneliness scale for large telephone surveys, adapted from the revised UCLA scale, asking how often people feel short of companionship, shut out or cut off from others. In a national sample of older adults, the short scale showed acceptable internal consistency and correlated strongly with the full scale, along with expected relationships with depression and social network measures.

Choosing Between Them

The three-item scale reduces burden and could be used for screening, but its limited range may make small changes harder to detect. I will use the three-item scale to screen residents for eligibility and the twenty-item scale at the start, after twelve weeks and again at twenty-four weeks as the primary outcome. Loneliness scores are treated as interval-level data for analysis, a common and generally acceptable practice for multi-item scales.

Secondary Measures

Loneliness is the main outcome, but two secondary measures will show whether any change spreads to mood and daily life.

Depressive symptoms will be measured with a brief screening scale validated in older adults. Participation in activities will be measured objectively from the community's activity sign-in records for the month before and the last month of each period. Call logs will record each call's date and length.

Practical Measurement Issues

Older residents may have vision limits, fatigue or difficulty with response formats. Questionnaires will be printed in large type, and a trained staff member blind to group assignment will offer to read items aloud. Administration will take place in residents' apartments or a quiet room, at times residents choose. I will record whether each questionnaire was self-completed or read aloud so that any effect of mode can be examined.

Levels of Measurement

The study's variables span several levels of measurement. Group assignment is nominal. Loneliness and depression scale totals are treated as interval. Activity attendance counts are ratio-level counts with a true zero. Demographic variables such as education are ordinal. Recognizing each variable's level guides the choice of statistics in the analysis plan, since counts, for example, may call for different models than scale scores.

Checking Measurement Quality

After baseline data collection, I will compute internal consistency for the loneliness scale in our sample and examine item responses for floor or ceiling effects. If the read-aloud and self-completed versions produce systematically different scores, I will note it as a limitation.

Conclusion

A census of eligible residents, a power analysis indicating about ninety participants, a validated loneliness scale with a short version for screening and accommodations for vision and fatigue give the study a sound foundation. Careful measurement planning now prevents uninterpretable results later.

References

Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175-191. https://doi.org/10.3758/BF03193146

Hughes, M. E., Waite, L. J., Hawkley, L. C., & Cacioppo, J. T. (2004). A short scale for measuring loneliness in large surveys: Results from two population-based studies. Research on Aging, 26(6), 655-672. https://doi.org/10.1177/0164027504268574

Russell, D. W. (1996). UCLA Loneliness Scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66(1), 20-40. https://doi.org/10.1207/s15327752jpa6601_2