Psychometric tests are often translated from one language into another for use in cultural and cross-cultural psychological research. This tradition has long been a primary method for studying psychological concepts across cultures. A body of literature has been emerging that raises concerns with this process of translation and suggests that psychological concepts may not always translate smoothly across cultures. In what follows, I examine these issues surrounding the translation of psychometric tests. To do this, I gathered survey data for an English translation of the Nunchi Scale, to examine the process of translation and various analytic techniques devised to demonstrate the validity and reliability of scales. The Nunchi Scale was developed by Korean researchers to measure the Korean psychosocial competency of nunchi (Heo & Park, 2013) which involves heightened awareness of others. Using ethnographic data and focus group data gathered by my team of Korean researchers, this paper critiques the psychometric translation process. In this paper, I will show that when this psychometric scale was translated from Korean to English, the very concept that was supposed to be measured by the scale, namely nunchi, was lost. From this process of translation, I will argue that the current practices of cross-cultural equivalence testing are not sufficient. This research offers a few cultural research methodologies to help check and correct against the issues that appear in translating or using foreign psychological concepts in cultural research. It is hoped that this research will help shed light on the complexities and problems involved in translating psychological concepts from one cultural context to another. It is also hoped that by understanding these complexities and problems, we will be able to better understand the psychological concepts that are purportedly measured by these psychometric measures. In sum, this paper offers a critical voice cautioning the use of untranslatable psychological concepts.

No existing measure to date captures mother-infant togetherness. A valid measure of togetherness is essential to engage in evidence-based practice, evaluate obstetric delivery models and nursing interventions, and measure the level of togetherness which promotes optimal maternal-infant outcomes. When together and in close proximity, a women and her infant have access to one another to allow for mutual caregiving or caregiving on cue. A new measure entitled the Mother-Infant Togetherness Scale (MITS) was developed based on a review of the literature and conceptual framework of Mother-Newborn Mutual Caregiving. The MITS is a 35- item instrument composed of four subscales that measures the timing, duration, and intensity of togetherness of the mother-infant dyad during entire hospitalization. The purpose of this multiphase study was to obtain support for the validity of the MITS. Phase 1 determined the content validity at the scale (S-CVI), subscale, and item level (ICVI) with a panel of expert judges. The final sample for the content validation consisted of 7 judges from medicine (n = 2), maternal-child nursing (n = 1), nursing research (n = 3), and social work (n = 1). Judges were instructed to use a 4-point Likert scale to rate the relevance of each item (I-CVI) to the construct of togetherness. The S-CVI was calculated from the mean I-CVI scores. The CVI-S of .88 was just slightly below the desired CVI-S ( > .90). Of the four subscales, all had adequate CVI ( > .90) at the subscale level except the delivery affective subscale (CVI = .74) and postpartum affective subscale (CVI = .89). The delivery events and postpartum events subscales had satisfactory CVI scores (CVI > .90), 1.00 and .94, respectively. The CVI-I results identified a total of seven items on the affective subscales that did not meet the desired I-CVI ( > .78). iv Phase 2 pre-tested the readability and understandability of the MITS among eight postpartum women. During the interviews, the women were asked to complete the MITS and provide opinions about the readability and understandability of the directions and items. The audiotapes were transcribed word for word, reviewed for thematic content, and revisions made to the study instrument accordingly. This same sample of postpartum women participated in the content validation of the delivery affective subscale (items #4a-j) and postpartum affective subscale (items #17a-j). The I-CVI results identified that a total of six items on the affective subscales had a CVI-I of .75, just slightly below the desired I-CVI ( > .78). Scale items were deleted or revised and the instrument retested until the desirable CVI at the scale and subscale level was achieved. Phase 3 used a descriptive study design to examine women's ability to accurately selfreport birth events on the MITS delivery events subscale at 4 weeks postpartum, as compared to observer-collected data obtained at delivery to determine the most valid mode of administration. A purposive sample consisted of 45 women having delivered at a community hospital in southwest Florida. The research team completed the MITS delivery events subscale immediately after delivery. Women were sent the MITS for completion 4 weeks after delivery. McNemar Chi-Squares were (χ) were calculated from the self-reported MITS delivery events subscale scores and the observer-collected MITS delivery events subscale scores. No significant difference (p < .05) was found supporting self-reported mode of administration for the MITS. Phase 4 is in-progress and evaluates the reliability and validity of the MITS subscale and total scale scores. The interim analysis was performed on a sample of 113 postpartum participants (composed of the final sample of 31 participants from Phase 3 and the first 82 participants from Phase 4) having delivered at three of the four participating hospital study sites. Adequate internal consistency reliability was found at the scale level with Cronbach’s alpha ( = .89) and split-half reliability results ( = .79 – 81, r = .83 - .88). Of the 35 MITS items, 10 items (28.6%) were found to have item-total correlations less than .30, arguing against treating MITS items as a single total scale measure. Good internal consistency was found at the delivery events subscale level (α = .78). Exploratory factor analysis (EFA) identified a twofactor solution. The two factors were named Taking In and Taking Control and had internal consistency reliability.79 and .65, respectively. Additional work needs to be done to improve the internal consistency of the Taking Control factor. The postpartum events subscale also had low internal consistency ( = .58). This subscale was not factor analyzed because the item response data did not meet the criteria for factor analysis. The items on the postpartum events subscale were assessed to be unique, singular, heterogeneous items that did not correlate well with other items. These results are conceptually logical given the nature of what the items are measuring (occurrence/intensity of specific events in time). The delivery affective subscale had good internal consistency reliability ( = .85) and a two factor solution. The two factors, named Feelings At Delivery and Delivery Concerns, had adequate internal constancy ( = .81 and  = .80, respectively). The postpartum affective subscale had good internal consistency reliability ( = .92) and a one factor solution. Results for known groups testing based on feeding type and mode of delivery found all group differences were in the predicted direction. Higher scores were found for mother-infant dyads who breastfed than for mother-infant dyads who bottle fed. However, only group differences for the events subscales were substantive and statistically significant (p < 001.). Higher scores were found for mother-infant dyads who experiencing a vaginal delivery than for mother-infant dyads who experienced a cesarean delivery. Group differences were substantive and statistically significant (p < .01) for three of the four subscale scores. A post hoc power analysis on the means and standard deviations from the interim analysis and the between-groups comparison effect size observed for feeding type (d = .50) found a sample of 45 adequate to have statistical power at the recommended beta of .80 and alpha of .05. The post hoc power analysis on the effect size for mode of delivery (d = .75), found a sample of 156 are needed to obtain statistical power at the recommended beta of .80 and alpha of .05. Therefore, the desired sample size of 200 women for the final analysis is adequate to obtain statistical power. A third known group testing for the variable of central nursery availability could not be performed with the interim analysis data because no participants in the interim analysis sample reported this experience. However, this analysis will be performed with the final data set. This is the first study to operationalize togetherness during the entire hospitalization and to include all dimensions of the construct. The findings from this multi-phase study provide initial support for the reliability and validity of the MITS. Although the results from Phase 4 are interim and therefore tentative, they provide preliminary psychometric evidence for construct validity.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.