Study Proposal: Developing Picture Naming Task /

Research Topic and Theoretical Background (this section is not clear at all … more info is needed)

Need to add a good intron here ….. Picture naming tasks are a common measure in neuropsychological assessments designed to assess aphasia, (Herbert et al., 2008) such as anomia (Randolph et al., 1999), as well as deficits of memory or executive deficits, such as those seen in dementia (Bayles & Tomoeda, 1983), and declines in normal aging (Zec et al., 2005). Need to add enough info her

A challenge of adapting currently existing tests is that picture naming tasks are psycholinguistic measures that are inherently language specific (Bates et al., 2003), and the items can also be culture-specific (Kremin et al., 2003), and may not be considered culturally fair. Even among the largest English speaking nations, there are different words for the US and UK versions of the Boston Naming Test (BNT) (Goodglass et al., 2001); one of the most frequently used picture naming tasks to assess aphasia (Harry & Crowe, 2014). Not clear!!!
Moreover, previous attempts to adapt the BNT, even to native English-speakers of New Zealand (Barker-Collo, 2001), and other languages (e.g., Spanish, Turkish) (Olabarrieta-Landa et al., 2015; Soylu & Cangöz, 2018), have resulted in changing up to half of the items in the original test in some instances (Soylu & Cangöz, 2018), which calls for commissioning new items of similar range of complexity and linguistic properties, but with items that are both linguistically valid and culturally fair.
The current study aims to match newly commissioned items, as well as items from a picture bank of psychometrically validated items (Rossion & Pourtois, 2004; Snodgrass & Vanderwart, 1980), after piloting all items, a selection will be made of culturally appropriate items that represent a range of complexity, and variety of linguistic properties, which will then be used in developing normative data. The final objective is to develop an English picture naming task, with a representative sample based on consensus data in terms of the following demographic factors: Age, Education, Region, Sex, and Socio-Economic Status (Alghamdi, 2017). ????

Stage 1: Item Selection (how about splitting the date into (positive negative snd neutral??? Any other important factors .. ??!!!
Cross cultural aspects that affect psychometric performance include psycholinguistic properties of both visual and verbal stimuli (Bates et al., 2003; Kremin et al., 2003). These properties can be separated into two categories of objective or subjective factors. Objective measures are intrinsic values of the words or the names of picture items, and include syllabic length, phoneme length, and orthographic frequency, i.e., how frequently they are presented in a lexicon (Barry et al., 1997).
Whereas subjective measures are based on ratings by participants of their impressions of a word, these include imageability (how easily an image is conjured of the term), image agreement (concurrence of the connotation between subjects), name agreement (the different nouns given by subjects), age of acquisition, familiarity, and subjective frequency (Boukadi et al., 2016; Khwaileh et al., 2018).
Additionally, the complexity of the image can be measured based on objective factors (e.g., pixel analysis), as well as subjective factors (Bates et al., 2003; Boukadi et al., 2016). Thus, this study will first collect objective measures of 108 commissioned line-drawings, as well as published data of English speakers of the Snodgrass picture bank (Rossion & Pourtois, 2004).

Stage 2: Pilot (not clear)
An initial selection will be made from the stimuli based on objective measures, and outliers, and items with redundant qualities will be eliminated. Then, these items will be presented to a pilot sample, and subjective ratings of the participants will be collected to then determine the final stimuli set. The sample size of the pilot study has been recommended to be a minimum of N = 30, and a cost-benefit analysis similarly indicates between 24 – 36 to be optimally sufficient, up to as many as 12 per group may be recruited for the sample size (Johanson & Brooks, 2010). How about having 3 groups?? Young adults, adults and older adults? Put the age range? And say why you did that? Why is it important?

Stage 3: Normative Data not clear …. Also you need to add the inclusion and exclusion criteria
After both steps 1) item selection, and 2) pilot study are completed, the final stimuli data that incorporate a representative sample of items based on psycholinguistic, and psychometric features will be compiled and administered to a large sample representative in terms of age, sex, educational levels, socioeconomic status.
Although sample sizes are inherently difficult to estimate a priori, because the final result will depend on whether the sample collected is representative of the entire population spectrum in terms of following a normal distribution (Crawford & Garthwaite, 2008). There are however a number of ways to estimate the sample size needed based on the number of groups and number of variables in question.
One such method is a cost-benefit analysis that shows that on a logarithmic scale, there are greater benefits for having 50-75 participants per group than having under 50 participants, but that value subscribes to the law of diminishing returns beyond N = 75, such that criteria as the confidence interval are not improved (i.e., narrower values) when increasing the N > 75 to justify the extra cost and time associated with recruiting that many more participants (Bridges & Holler, 2007).
As sample sizes used for the BNT are deemed poor and lacking especially in older populations (Harry & Crowe, 2014; Kent & Luszcz, 2002), we seek to emulate a large sample adaptation of the BNT on the Spanish speaking population of Latin America, who had the adult populations segmented in 13 groups, each group spanning 5-year intervals from 18 years to 77+ (Olabarrieta-Landa et al., 2015). Following the cost-efficient recommendation of sample size to establish normative data (Bridges & Holler, 2007), we seek to recruit 50-75 per age group, resulting in 650 – 975 participants in total. Following data collection, a confirmatory factorial analysis will be conducted to establish the normative data, and a subsequent validation, and reliability ..