Rasmussen College Essentials of Social Work Measurement Discussion
Response #11) Suppose a geriatric social worker assesses whether a life history review intervention improves the level of depression among frail nursing home residents by administering a depression measure to them before and after the intervention. Suppose the measure had its validity assessed by comparing scores of frail nursing home residents on it to the scores of healthy elderly folks living independently. What type (and subtype) of validity was assessed? What more would be needed to establish the measure's construct validity? If the measure is valid, can we assume it is also reliable? Why?When the social worker compared the results of frail nursing home residents to healthy elderly folks living independently, he was testing the known-group criterion validity. Rubin and Babbie (2016, pg. 204) explain that criterion validity measures validity by comparing results to an external criterion. They further explain that known-groups criterion validity measurements are based on groups the researcher expects will have predetermined results (Rubin and Babbie, 2016). In this example, the researcher expects elders who are healthy and living independently will have lower scores on a depression scale than the nursing home residents, which will support the validity of the scale.In order to better establish the measure’s construct validity, the researcher could look at diagnoses of the nursing home residents and the elderly living independently. This could only be done for those individuals who have spoken to their doctor or have seen a mental health professional, but it would allow the researcher to compare his results with the diagnoses, or lack thereof, from the doctors and mental health professionals. If the higher scores on the scale also matched the individuals with a previous diagnosis of depression, the researcher would have evidence supporting the construct validity of his scale. Rubin and Babbie (2016, pg. 206) describe this same process through the example of marital satisfaction. In that case, a scale of marital satisfaction would be compared to clinician’s records regarding marital satisfaction.If the measure is valid, we can assume it also is reliable. As Dr. Melissa Radey (2020) explains in her PowerPoint and lecture notes, a measure can be reliable and invalid, but it cannot be valid and unreliable.2) A geriatric social worker develops a very lengthy scale to assess the degree of staff sensitivity to the needs of residents in nursing homes. The scale is designed to be completed by both staff and residents. Identify reasons why the scale might be vulnerable to both systematic and random error. Describe two ways to test the reliability of the scale.Rubin and Babbie (2016, pg. 192) explain that systematic error is an error that is reliable. The error can be replicated and will be replicated in each additional attempt. One reason the scale might be vulnerable to systematic error could be that the staff members want to be viewed favorably, so they give the answers they believe to be favorable, regardless of whether those answers are true. Additionally, some residents might not understand the questions, so they might just give the same answers to every question, regardless of what the question asks.However, that same problem could instead result in a random error because the residents might simply guess at answers if they don’t understand the questions. And if the survey were written with the staff members in mind, it’s quite possible that older residents might not understand. Rubin and Babbie (2016, pg. 195) explain random error as an unpredictable error – an error that is not consistent. In this example, the staff members might give random answers if they feel rushed or don’t believe they have the time to spare the survey. That also would result in a random error to the measurement.The reliability of the scale could be tested through the test-retest method. Rubin and Babbie (2016, pg. 201) explain that the test-retest method can be done about two weeks after the first survey for the best results. That way, the subjects likely won’t have time to have changed, thereby altering their responses. Using this method, it should be possible to weed out any random error. Those who selected random responses the first time are likely to do it again, and probability nearly guarantees that they will select different random answers the second time.Additionally, the researcher could test the internal consistency reliability of the survey, perhaps by splitting it down the middle and using software to calculate the coefficient alpha, as described by Rubin and Babbie (2016, pg. 201). This likely would prove the test to be unreliable because the test is attempting to measure staff sensitivity by questioning two groups with one test. It seems to me that the researcher needs two surveys for such a study because the staff and residents will need different questions. Understanding that, I expect testing the internal consistency reliability would prove that the test is attempting to test two different things in one survey.Radey, M. (2020). Week 6 (Measurement). [PowerPoint slides]. Canvas.fsu.edu. https://cas.fsu.edu/cas/loginRubin, A. & Babbie, E. (2016). Research methods for social work. Cengage Learning.Response #21. Suppose a geriatric social worker assesses whether a life history review intervention improves the level of depression among frail nursing home residents by administering a depression measure to them before and after the intervention. Suppose the measure had its validity assessed by comparing scores of frail nursing home residents on it to the scores of healthy elderly folks living independently. What type (and subtype) of validity was assessed? What more would be needed to establish the measure's construct validity? If the measure is valid, can we assume it is also reliable? Why? Validity measured by Comparing scores of frail nursing home residents to healthy independant elderlyWhat type & subtype of validity was measured? Type of validity used to measure the variable (Depression):Convergent validity Subtype: Concurrent validity Initially, I thought the two comparisons were not related due to the living situations and associated health issues or lack of health issues of each cohort. Then I thought more about convergent validity. Convergent validity removes the relationship aspect of construct validity but compares the variable in two separate situations (Radey, 2020, slide 14). Therefore, it turns the validation into depression in A vs. depression in B, comparing the two.What more would be needed to establish the measure's construct validity? Construct validity translated to; does the test measure what it is supposed to measure? According to the text, construct validity measures variables that are similar (Rubin & Babbie, 2017, p.206). This helps to differentiate the perception of a variable an individual may have when completing a survey. To measure the construct validity of depression in this age group, a researcher, could ask about: The feeling of worthlessness or hopelessness Sadness or lack of interest in activities The feeling of wanting to die or commit suicide If the measure is valid, can we assume it is also reliable? Why? If a measure is valid, it is also reliable (Radey, 2020, slide 15). Why? To be reliable means to get the same results repeatedly; to be valid is to get the right results over and over again (Radey, 2020, Slide 15). As an example, food item nutrition data may be reliable but lack content validity. Content validity is not just whether what is measured is correct; it also includes the factors or perceptions that make up the concept (Rubin & Babbie, 2017, p.203). Therefore, just because a food item is measured over and over again and reflects the same .45 grams of sugar per serving deeming the item sugar free according to the FDA guidelines (U.S Food & Drug Administration, 2007). This measurement lacks content validity because it is not sugar-free. Some people may believe that there is absolutely no sugar per serving and, therefore, no sugar in the container, even though technically there are 18.5 grams of sugar in a container; which is almost equal to ½ of a 12 ounce can of coke (The coca-cola Company, 2020).2. A geriatric social workers develops a very lengthy scale to assess the degree of staff sensitivity to the needs of residents in nursing homes. The scale is designed to be completed by both staff and residents. Identify reasons why the scale might be vulnerable to both systematic and random error. Describe two ways to test the reliability of the scaleIdentify reasons why the scale might be vulnerable to both systematic and random error One reason the measured scale is vulnerable is due to social desirability bias on the part of the staff, which is when people act or do things to make either themselves or the group they are a part of look good (Rubin & Babbie, 2017, p.193). Logically, they would want to portray themselves in the best light they can. Staff would be indirectly communicating with their employer, focusing a negative light on themselves or others in their unit, which would not be beneficial to their potential employment with the agency. Additionally, the residents may not want to reflect their real views, fearing that they may have issues with the staff taking care of them due to the information they provided. Another bias that may lead to error is the acquiescence response set. Some individuals simply go along with the statements, no matter what is said (Rubin & Babbie, 2017, p. 193). That being said, these people may not contemplate what they are being asked or reflect on their experiences with the staff. Describe two ways to test the reliability of the scaleOne method of testing the reliability of a scale is to do a double check, with the same group at a different time; if the results match up at the very least a rate of 70 percent, then the scale is deemed reliability (Rubin & Babbie, 2017, p.201).Another way to test the reliability of a scale is to check the internal consistency reliability. Testing this way involves testing the subsets and matching the results of all the subsets to the total score earlier notes for the entire scale Rubin & Babbie, 2017, p.201).ReferencesRadey, M. (2020, June 15). Week 6 (Measurement).ppt. Retrieved from canvas.fsu.edu: https://canvas.fsu.edu/courses/128586/files/823798...Rubin, A., & Babbie, E. R. (2017). Research Methods for social Work. Boston: Cegage Learning.The coca-cola Company. (2020). How much sugar is in a coca-cola? Retrieved from coca-colacompany.com: https://www.coca-colacompany.com/faqs/how-much-sugar-is-in-coca-cola#:~:text=There%20are%2039%20grams%20of,12%20oz%20Coca%2DCola%20can.U.S Food & Drug Administration. (2007, September). Guidance for Industry and FDA: Dear Manufacturer Letter Regarding Sugar Free Claims. Retrieved from fda.gov: https://www.fda.gov/regulatory-information/search-...