PHE 327 Module 6 Analysis Short Paper Example

What 548 Surveys Say: An Analysis of Loneliness and Housebound Winters Among the County's Seniors

[Student Name]

Southern New Hampshire University

PHE 327: Research and Assessment in Public Health

Module Six Short Paper

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

What 548 Surveys Say: An Analysis of Loneliness and Housebound Winters Among the County's Seniors

The survey of older adults described in Project One went out in December and January. This paper analyzes what came back. It reports the response, describes the main results with their uncertainty, makes two planned comparisons, explains how weighting changed the estimates and looks separately at the smaller group reached through meal drivers and mail carriers. Interpretation for action is left to the assessment report in Project Two. Every figure below can be traced to the returned forms and the weighting file kept with the data.

Response

Of 1,200 households sampled, 548 returned a completed survey, a response rate of 45.7%. Most came back by mail after the second or third contact, and 41 were completed by phone with a senior center volunteer. Dillman et al. (2014) describe how multiple mailings and alternative response modes raise participation, and the pattern here fits: the replacement survey alone added 96 responses. Respondents were somewhat older and more often women than the county's older population as a whole, and slightly fewer lived alone than census figures suggest, a pattern consistent with the concern that isolated people respond less. Phone responses came mostly from people over 80.

Main Results

Table 1 summarizes the main indicators for the 548 respondents before weighting. On the three-item scale of Hughes et al. (2004), totals run from three to nine, and anyone scoring six or above is counted as lonely.

Table 1. Main Survey Results, Unweighted (n = 548)

Note. Confidence intervals use the normal approximation for a proportion.

Reading the Intervals

The confidence intervals express uncertainty from sampling. The loneliness estimate of 31% means that, if the sample is representative, the true share among older adults on the sampling list is likely between about 27% and 35%. The intervals do not account for nonresponse bias or measurement error, which could make the true figures higher. The reliability of the loneliness scale in these data was acceptable, with a coefficient of 0.81, similar to what its developers reported, which supports using the score as intended.

Comparisons

Two comparisons were planned. First, loneliness by living arrangement: 92 of 210 people living alone were lonely, 43.8%, compared with 78 of 338 people living with others, 23.1%. A chi-square test showed this difference was very unlikely to be due to chance (p < 0.001). Second, confinement by area: 41.3% of remote-township respondents were housebound a week or longer, against 36.6% in the three towns. This difference was not statistically significant (p = 0.31), so the survey does not show that remote residents are more often confined, even though partners expected they would be. Reporting a non-significant result plainly matters, because it guards against building a program on an assumption the data do not support.

Weighting

Because people living alone responded less than expected, results were weighted to match census figures for age group, sex and living alone. Weighting raised the estimated share who were lonely from 31.0% to 32.4% and the share housebound a week or longer from 38.0% to 39.1%. The changes are modest but move in the expected direction: correcting for underrepresented isolated people raises estimates of isolation. Weighting cannot correct for differences that were not measured, so the true figures may be higher still.

The Separate Driver and Carrier Group

Meal drivers and mail carriers reached 61 older adults living alone who agreed to answer a short version. Of these, 52% were lonely and 61% had spent a week or more without leaving home. This group is not a random sample and cannot be combined with the main estimate, but its much higher figures suggest that the mailed survey missed some of the most isolated residents. Holt-Lunstad et al. (2015) found isolation and loneliness linked to higher mortality, which makes these missed residents a priority even though they cannot be counted precisely.

Open-Ended Answers

Of the respondents, 312 wrote something in the open question about what would help. The most common themes were rides to appointments and the grocery store in winter, a regular phone call or visit, help with snow removal and lower heating costs. Several wrote that they did not want to be a burden, which may explain why some who are isolated do not ask for help.

Checking the New Items

Module Three promised two simple checks on the new questions. Both came out as expected. Respondents who spent a week or more housebound had higher loneliness scores on average than those who did not, 5.9 compared with 4.6, which suggests the confinement item captures something related to isolation rather than simply bad weather. Missed medical appointments were more common among respondents living more than twenty miles from the nearest clinic, 34% against 22%, which fits the item's intended meaning. The heat-or-necessities question showed the expected link with low income on the census-matched variables available. These checks do not prove the items are perfect, but they give no sign that the new questions misled respondents.

Item nonresponse was low for most questions, under 3%, but higher for the heat question, 7%, which may reflect discomfort with discussing money. Those missing answers were excluded from that indicator only, and the report will note that the true figure could be somewhat higher.

Conclusion

About a third of older adults in the county report loneliness and nearly two in five spent a week or more indoors last winter, with loneliness much more common among those living alone. Remote residents were not clearly more confined than town residents. The hardest-to-reach group appears worse off than the main sample. Project Two will turn these findings into an assessment of needs, assets and capacity. Those figures are estimates with clear limits, and the report will present them that way.

References

Dillman, D. A., Smyth, J. D., & Christian, L. M. (2014). Internet, phone, mail, and mixed-mode surveys: The tailored design method (4th ed.). Wiley.

Holt-Lunstad, J., Smith, T. B., Baker, M., Harris, T., & Stephenson, D. (2015). Loneliness and social isolation as risk factors for mortality: A meta-analytic review. Perspectives on Psychological Science, 10(2), 227-237. https://doi.org/10.1177/1745691614568352

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