It looks like you're referring to a needs assessment, since you're interested in determining whether a phenomenon exists and describing it. In several variations of that design, evaluators are looking to determine: 1) is there a problem that warrants an intervention? 2) What does that problem look like? 3) How should an intervention be formulated to make the services/program/policy/intervention useable to those they are intended to support?

This problem identification could involve a survey (as you bring up), an epidemiological study of the problem and its rates in the community/geographic region, etc. Surveys are a legitimate option, but it's probably helpful to incorporate some interviews or a focus group depending on what problem you're examining (e.g., how common or rare you suspect the problem is).

Here is a well-cited book about this methodology that I think can point you in the right direction:

Gupta, K. (2011). A practical guide to needs assessment. John Wiley & Sons.

Thank you William for your comments, suggestions and book references. I think this will be very helpful. For additional information, in the introduction/background of this empirical research problem I am collecting data evidence through preliminary surveys about the existence of real practical problems (problems faced by company employees related to people and organization studies, for example: satisfaction, performance, motivation, etc.). I hope to get preliminary empirical data that I collect myself that it is true that there are practical problems in the object I am researching. I need a preliminary survey literature source (not a pre test or pilot test because to my knowledge these two things are not in the introduction section but methodology) as a scientific guideline on how many respondents are statistically feasible, what are the procedures for minimising bias, processed using what type of statistics? (can I just use simple descriptive statistics?).

In your case, the answer is probably, "yes." You're just trying to get a baseline for those metrics. Points of central tendency (i.e., mean, median mode) are perfectly fine descriptors of the data as long as you're paying attention to factors like skewness and kurtosis. Select the right option based on the distribution you're seeing when you analyze the data.

For missingness, this depends on whether you're assessing this problem for one agency/company or if you're trying to generalize to the industry. If you're generalizing to the industry, then I highly recommend this book to help you with sampling and survey design before getting into that:

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

After you've worked through your sampling strategy, it will help to try and determine whether your data are missing at random (MAR), missing completely at random (MCAR), or missing not at random (MNAR). Look into power for imputation strategies if that is of interest. I would only bother with imputation strategies if I were trying to design a representative survey of the population in the industry--not a single company or smaller group.

Thank you William T. Miller for taking the time to provide complete answers and explanations. The research I am doing now is still on a smaller scope of research that is internal to one company/organization. Suppose I need to get the questionnaire items written in the book you have given before. What is the solution for the questionnaire items of a construct as a basis for conducting a preliminary survey?

I am only interested in learning more about missing at random (MAR), missing completely at random (MCAR), or missing not at random (MNAR) and imputation strategies related to practical problems in an industry. I would appreciate suggestions for literature sources that I can learn from.

The book is intended to help you with questionnaire construction and with sampling/solicitation requests--it won't have many items for you to adopt.

In the case of an exploratory study, I highly recommend general questions (e.g., satisfaction with culture, satisfaction with performance ratings, perceptions of performance, self-reported motivation) and having open-ended questions that gather more insight based on the responses. For example, you could have a survey that incorporates logic to show the respondent the open-ended question if they report being unmotivated or highly unmotivated. Add in some open-ended questions that gather general feedback you can explore later. For example, "If you could change one thing about XXXX, what would it be?" Your HR department would be a great resource for determining the type(s) of concerns they feel are worth addressing. Otherwise, look into the HR literature for factors that predict attrition or whatever concept you're interested in exploring down the road.

As far as drafting questions goes, map out your concepts first. Create an exhausting list of all the topics you want to cover. Rank them. Find Likert scales for them (e.g., strongly satisfied - strongly dissatisfied; very motivated - very unmotivated). Develop a survey that takes less than 10 minutes after testing with someone informally.

Here's a resource on missing data for starters:

9.2 MCAR, MAR, MNAR | Introduction to Regression Methods for Public Health Using R (bookdown.org)

Thank you, William T. Miller. An exploratory study can be used in conducting a preliminary survey to find empirical data evidence related to the problems experienced by company employees. What are the literature sources that support this? And what is the minimum number of respondents? Is it necessary to do validity and reliability tests? Because if you survey with open-ended questions, of course, after the data is collected, it cannot be processed with descriptive statistics.

Astadi Pangarso, regarding the minimum number of respondents: These are easy to gather if you're doing a power analysis for a specific statistical analysis, but since you're going to be doing basic descriptive statistics, this is much easier. A sample size of 30 is widely accepted as a minimum number of respondents for these analyses, but you're better off determining whether responding was associated with any variable of interest and using that as your baseline. For example, if you notice that you have most of your responses from an accounting department, you need to reach out to other departments. If you notice mainly women responded, determine whether that's proportionate to your company's population. You don't want to focus on minimum numbers of responses, since it's not like you won't have statistical power. You should be far more concerned about generalizability. Don't worry so much about minimum numbers and focus instead on whether you have responses representing the company. You should be happy with 30 - 50. Citation for this argument and a book I recommend:

Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R. (2009). Survey methodology. Wiley.

Reliability and validity aren't as important here as many may assume. The reason is that you're going to be reporting descriptive statistics. If you were building scales and intending to use them in another analysis or to share the scales with the public for widespread use, then psychometrics comes more into play. If you're not going to report based on scale values (e.g., summated scales or factor analysis-generated measures), it's overthinking the purpose of the project. The book below outlines how psychometrics is intended to work.

Raykov, T., & Marcoulides, G. A. (2010). Introduction to psychometric theory. Routledge.

Open-ended questions are going to be checked for themes. You're going to identify patterns of responses. What do people mention most often? In what context? You may even just dump all of the responses and display them directly if you have very few that warrant attention (i.e., 5-10). Pull some useful quotes that best fit the theme. Explore Journal of Qualitative Research for some insight into common practices.

Thank you William. T. Miller, your complete answer is beneficial to me; I wish you all the best...is there any possibility of collaborative research and manuscript writing for research publication in international journals?

Absolutely. My recent research focuses on criminal justice occupational training and testing, so vocational studies could be an opportunity for collaboration.