Levels of Measurement
In data collection, levels of measurement are essential because they determine the quantitative analysis of data to obtain descriptive statistics. Given that the study aims to use a descriptive research design, the collection of data should allow quantitative analysis data. Miller and Yang (2007) advise researchers to be careful when planning data collection because the levels of measurement that they use determine the nature of quantitative analyses that they perform. In this view, the levels of measurement should allow an analysis of data using descriptive statistics. As the study uses sources of funds as an independent variable, the appropriate level of measurement is a nominal scale. A nominal scale is appropriate in measuring the existence of two or more categories of a given variable (Stangor, 2010). Hence, as sources of funds are many, a nominal scale is suitable for measuring them.
In the measurement of the independent variable, an interval or numerical scale is appropriate. The independent variable in the study is the financial sustainability of the library given the sources of funds. As that financial sustainability is subjective, the study will transform the numerical scale into an ordinal scale in the form of a Likert scale. Likert scale is a measurement scale that depicts the degree of a certain variable based on certain intervals, which are in an ordinal manner (Royse, 2008). In this view, the study will categorize financial sustainability into three-Likert scale items, namely, unsustainable, low sustainability, and high sustainability. The sources of funds can fall in these three categories of financial sustainability, and thus provide their impacts on the sustainability of Clayton County Library System is offering its services. Babbie (2012) states that the Likert scale converts numerical data into ordinal scale, and thus depicts how the distribution of data occurs. Therefore, the ordinal scale in the form of the Liker scale is an appropriate level of measurement that the study will utilize in measuring the dependent variable, financial sustainability.
Validity is central in research because it assures the accuracy of data that a certain instrument measures. Essentially, validity refers to the ability of a research instrument to measure a given variable with accuracy while eliminating the existence of confounding variables (Thyer, 2010). Normally, there are three types of validity, viz. content validity, empirical validity, and construct validity (McBurney & White, 2009). Content validity is the ability of the research instrument to cover all variables of a given phenomenon. In this case, since the study seeks to measure the influence of sources of funds on the financial sustainability of the library, it will enhance the content validity of the study by analyzing all sources of funds and classifying their sustainability using an accurate Likert scale. The dependent and independent variables should have causal relationships to enhance content validity (Stevens, 2006). Thus, the independent and dependent variables must cover all factors that mediate their relationships.
Empirical validity is a critical form of validity because it measures the ability of a research instrument to provide empirical outcomes. According to Groves (2005), empirical validity is the extent to which the dependent variable correlates with an observable variable. In this view, the financial sustainability of the library (dependent variable) must correlate with the quality of services that the library offers (observable variable). Thus, to increase the empirical validity, the study should ensure that the categorization of funds into unsustainable, low sustainability, and high sustainability is accurate enough to provide the empirical application. Empirical validity allows the prediction of outcomes, and thus forms the theoretical basis of research (Connaway & Powell, 2010). Predictably, diverse sources of funds would enhance the financial sustainability of the library. Moreover, a great deal of funds guarantees the financial sustainability of the library. Accurate estimation of funds necessary to enhance the sustainability of the library would increase the empirical validity of the study.
Construct validity measures the relationships between independent and dependent variables. Construct validity seeks to determine if a given construct can accurately predict the trend of another construct (Goldenberg, 2008). A construct that can effectively predict another construct has a high level of construct validity, while a construct that partially predicts another construct has a low level of construct validity. In the case study, the sources of funds should predict the financial sustainability of the library. To enhance construct validity, the study will ensure that it considers major sources of funds, which the library relies on providing quality services to the clients. Mackey and Gass (2005) argue that the use of multiple estimates of a certain construct would enhance construct validity. Fundamentally, sources of funds should be diverse while prediction of financial sustainability of the library should be reasonable.
Reliability of Measurement
Reliability of measures refers to the consistency of the findings when measured at different times. Meyer (2010) defines reliability as “the degree of the test score consistency over many replications of a test or performance of a task” (p. 4). The measurement is reliable if it provides the same findings, which are independent of time. To enhance the reliability of the survey, the study will design a questionnaire that measures all sources of funds. Since sources of funds form an independent variable, they determine the accuracy of the data that the study collects, and consequently reliability. In the development of questionnaires, the questions should not be ambiguous to enhance the consistency of responses (Andrew, Pedersen, & McRvoy, 2011). Therefore, the study should ensure that the questionnaire focuses on all aspects of the independent variable (sources of funds) to enhance precision. Since the independent variable predicts the trend of the dependent variable, its development in the questionnaire should be reliable.
Moreover, the dependent variable is a significant predictor of the reliability of the questionnaire. Given that financial sustainability exists on an ordinal scale that measures the degree of sustainability in the form of a Likert Scale, the design of the Likert scale affects the reliability of the instrument. The design of a Likert scale determines the reliability of the questions in the questionnaire (Brace, 2008). A poorly designed Likert scaled questions reduce the reliability of the research instrument. Thus, to enhance the reliability of the questions in the Likert scale, the study will use a three-item Likert scale, unsustainable, low sustainability, and high sustainability. A three-item Likert scale is more reliable than a five-item Likert scale because respondents find it easy when providing their responses (Lyberg, Biemer, Collins, Leeuw, Dippo, Schwarz, & Trewin, 2012). Respondents prefer Likert scales that have fewer items than those with many items because they are not confusing or cumbersome. The complexity of the Likert scale items and responses usually reduces the reliability of the research instrument. In this view, the study will ensure that a three-item Likert scale is designed properly to cover all aspects of financial sustainability, and thus enhance the reliability of the questions.
Strengths and Limitations of the Measurement Instrument
A questionnaire is an important research instrument that researchers normally use when collecting data from respondents. Regarding validity and reliability, the questionnaire has several strengths. Like the strength, a questionnaire has a strong internal validity because a researcher designs it to measure a given phenomenon. A researcher enhances the internal validity of the questionnaire by ensuring that construct validity and content validity are high. According to Elmes, Kantowitz, and Roediger (2011), the ability of researchers to manipulate questionnaires in response to the unique needs of participants is central in enhancing the external validity of a study. The simplicity of the questionnaire enhances internal validity because respondents would hardly provide wrong information owing to confusion (Lyberg, Biemer, Collins, Leeuw, Dippo, Schwarz, & Trewin, 2012). Thus, the ability of researchers to design and manipulate questionnaires increases the internal validity of the questionnaire.
A questionnaire also has external validity, which is a strength. Since the questionnaire is easy to administer to a large number of participants at once, it increases the external validity of the findings. A large number of participants promote the external validity of the findings, and thus generalizability. Concerning reliability, a questionnaire is a very reliable instrument, which is the strength, because it measures the same variable among various participants. Peterson (2000) argues that questionnaires are reliable because they are independent of researchers’ influence during the process of data collection. Comparatively, in interviews, researchers tend to ask leading questions in a bid to influence the responses of participants (Kumar, 2010). Hence, questionnaires are very reliable instruments because they experience minimal biases during data collection. Moreover, recoding numerical data into a three-item Likert scale enhances the reliability of the questionnaire. Gravetter and Forzano (2010) affirm that a structured questionnaire in the form of a Likert scale is more reliable than an unstructured questionnaire. Therefore, the structuring of questions in the form of the Likert scale is the strength of the questionnaire in this study.
Despite having numerous strengths, the use of questionnaires also has some limitations. The first limitation is that the data, which questionnaires collect, may be inaccurate, and thus has low internal validity. The validity of the findings or collected information is dependent on the veracity of the information that respondents provide (Rea, & Parker, 2005). By using questionnaires, it is difficult to establish the accuracy of the information and the sincerity of the participants. Normally, the questionnaire is valid under the assumption that the respondents are honest and sincere when providing their responses. The second limitation is that the structuring of questionnaires normally overlooks other important factors, which may be having a significant influence on an independent variable (Little, 2013). Hence, the internal validity of structured questionnaires in the form of Likert scales is dependent on the researchers. Regarding reliability, selection bias is a limitation that reduces the reliability of the questionnaire. Treiman (2009) holds that selection bias is a limitation of questionnaires because it entails the selection of participants or objects of study. The selection varies from one researcher to another depending on the biases associated with sampling.
The appropriate scale that the study requires to measure the independent variable (financial sustainability) is a numerical scale. Quantitative research uses numerical data, which researchers can easily analyze to provide descriptive statistics. In this case, the sustainability of the library correlates with the number of financial resources that the library receives from various donors. Hence, analysis of the financial resources requires the application of numerical scale. Singh (2007) affirms that quantitative data require numerical analysis of data to provide robust statistical analyses, which anyone can interpret quite easily. Therefore, since the research examines quantitative data in terms of financial resources, a numerical scale is appropriate in data collection. Furthermore, the study should convert the independent variable into a Likert scale to enhance the analysis of the numerical data. The likert scale provides the degree of variation of a certain variable (Thyer, 2010). In this view, financial sustainability varies from unsustainable to high sustainability. Conversion of the numerical variable into ordinal variable categorizes different sources of funds into different categories, unsustainable, low sustainability, and high sustainability. Eventually, the Likert scale is the most appropriate in representing the financial sustainability of the library relative to various sources of funds.
Justifying Reliability and Validity of the Scale
To justify the validity and reliability of the Likert scale, the pilot study is essential. The study will select a well-established library that receives sources of funds from various sources and is promising to be financially sustainable. Stangor (2010) argues that pilot testing allows researchers to test the validity and reliability of their instruments before performing the actual study. To test validity, the study will use the Likert scale in establishing if it measures the financial sustainability of a well-established library, which is very sustainable. If the scale provides predictable findings, then it means it is valid. Regarding reliability, the study will instruct several researchers to administer the questionnaire to various people at different times. The consistency of the findings obtained will determine if the Likert scale is accurate or not. Thus, pilot testing is the best way of demonstrating the validity and reliability of a research instrument or scale.
The appropriate test for the study plan is the norm-referenced test. The norm-referenced test compares the performance of individuals or entities with a view of ranking them to establish best performers and worst performers (Ary, Jacobs, Razavieh, & Sorensen, 2010). In this case, the study seeks to establish the financial sustainability of the library based on its performance among other libraries. If the library ranks above average, it means that its performance is sustainable, but if it ranks below average, it means that its performance is unsustainable. However, the cutoff point of sustainable and unsustainable performance is very subjective depending on the discretion of the researchers. A norm-referenced test is appropriate for the study because numerous factors determine the sustainability of the library. In essence, the criterion-referenced test cannot capture all the factors that determine the sustainability of the library, and thus makes the norm-referenced test an alternative test that is appropriate.
Population Used for the Scale and Test
The populations that the study will use are librarians and library users. Given that librarians understand the management of libraries and the nature of services that they offer, they are critical participants in the study of the financial sustainability of Clayton County Public Library. According to Daugherty and Russo (2013) librarians plays a central role in the management of a library and consequently determine its sustainability. Hence, as librarians are very knowledgeable about the management and nature of services they provide to the public, they are resourceful to the study. In addition, library users comprise an important population of study because they understand the quality of services they receive from the library. In this view, they can provide an invaluable view that can enable researchers to predict the sustainability of the Clayton County Public Library.
Andrew, D., Pedersen, P., & McRvoy, C. (2011). Research methods and design in sport management. London: Human Kinetics.
Ary, D., Jacobs, L., Razavieh, A., & Sorensen, C. (2010). Introduction to research in education. New York: Cengage Learning.
Babbie, E. (2012). The basics of social research. New York: Cengage Learning.
Brace, I. (2008). Questionnaire design: how to plan, structure and write survey material for effective market research. New York: Kogan Page Publishers.
Connaway, L., & Powell, R. (2010). Basic research methods for librarians. London: ABC-CLIO.
Daugherty, A., & Russo, M. (2013). Embedded librarianship: what every academic librarian should know. London: ABC-CLIO.
Elmes, D., Kantowitz, B., & Roediger, H. (2011). Research methods in psychology. New York: Cengage Learning.
Goldenberg, R. (2008). Video-based situational interviewing: construct validity, group differences, and utility. Washington: The George Washington University.
Gravetter, F., & Forzano, L. (2010). Research methods for the behavioral sciences. New York: Cengage Learning.
Groves, R. (2005). Survey errors and survey costs. London: Johns Wiley & Sons.
Kumar, R. (2010). Research Methodology: A step-by-step guide for beginners. New York: SAGE Publisher.
Little, D. (2013). The Oxford handbook of quantitative methods in psychology. New York: Oxford University Press.
Lyberg, L., Biemer, P., Collins, M., Leeuw, E., Dippo, C., Schwarz, N., & Trewin, D. (2012). Survey measurement and process quality. London: John Wiley & Sons.
Mackey, A., & Gass, S. (2005). Second language research: Methodology and design. New York: Routledge Publisher:
McBurney, D., & White, T. (2009). Research methods. London: Cengage Learning.
Meyer, P. (2010). Understanding measurement: reliability. Oxford University Press.
Miller, G., & Yang, K. (2007). Handbook of research methods in public administration. New York: CRC Press.
Peterson, R. (2000). Constructing effective questionnaires. London: SAGE Publisher.
Rea, L., & Parker, R. (2005). Designing and conducting a survey research: A comprehensive guide. New York: Oxford University Press.
Royse, D. (2008). Research methods in social research. New York: Cengage Learning.
Singh, K. (2007). Quantitative social research methods. London: SAGE Publisher.
Stangor, C. (2010). Research methods for the behavioral sciences. New York: Cengage Learning.
Stevens, S. (2006). An investigation of the content validity, stability, and internal consistency of the Spanish version of the transition planning inventory home form. Kansas: University of Kansas.
Thyer, B. (2010). The handbook of social work research Methods. London: SAGE Publisher.
Treiman, D. (2009). Quantitative data analysis: Doing social research to test ideas. London: John Wiley & Sons.