Sampling Concepts PDF

SH1806 Ways and Means of Collecting Data I. Sampling Concepts Sampling is a method of acquiring representatives of a certain population to gain and determine parameters of the whole group (Merriam-Webster, 2018). The term population is a term used in research...

SH1806 Ways and Means of Collecting Data I. Sampling Concepts Sampling is a method of acquiring representatives of a certain population to gain and determine parameters of the whole group (Merriam-Webster, 2018). The term population is a term used in research that denotes a huge group of people where you can choose a sample. This sample represents the entire group. Sampling frame is the list of the members of the population to which the researcher wants to generalize or apply his or her findings about the sample. The sample unit denotes the individuals in the population. For example: Population STI Senior High School students from Metro Manila Sample 1,500 STI Senior High School students from Metro Manila Sampling Frame All STI Senior High School students from Metro Manila who attended public schools Sampling Unit All male STI Senior High School students from Metro Manila who attended public schools Researchers must be aware that there are factors that can affect sample selection. According to Babbie (2016), when selecting respondents for a study, a researcher must not only listen to his/her mind but by other factors like the ones listed below: 1. Sample Size – it refers to how small or big the sample size is. A few researchers would often base their sample size on studies, which they have previously read. The best way to determine the appropriate sample size is the representativeness of the sample with respect to the population. This can be achieved through randomization of samples. Through randomization, the chances of acquiring a 95% or higher distribution of the population is highly probable because it will target all aspects of the population unbiased of economic status, sex, educational attainment, and others. Researchers must attain a confidence level or the.05 level since it is the acceptable degree of representativeness of samples. 2. Sampling Technique – this factor falls under two (2) categories: a) probability sampling which is bias-free due to the use of randomizations; and b) non-probability sampling which uses pre- selected samples and is prone to bias. Bias is one of the leading factors that can cause sampling errors. 3. Heterogeneity of the Population – heterogeneous population is composed of individuals with varied capabilities and characteristics. It is important to choose a sampling technique that will widely distribute the selection of a large sample among all members of the population. This will cover each nook and cranny of the demographic and will prevent sampling errors. 4. Statistical Techniques – the accuracy and exactness of the samples depend on how the researcher effectively used his/her mathematical methods, in this case, statistics. Errors in using mathematical methods could yield negative and erroneous results that could alter the whole study. 5. Time and Cost – covering a sample within a population would need a lot of time and money. One good example is the Food and Nutrition Research Institute Department of Science and Technology (FNRI-DOST) and their research program called the National Nutrition Survey, which aimed in monitoring the nutritional status of Filipinos from Batanes to Tawi-Tawi which 06 Handout 1 *Property of STI [email protected] Page 1 of 5  SH1806 is the Southernmost part of the Philippines. Personnel from this government agency have roughly five (5) years to gather data from all over the Philippines and get a huge chunk of budget from the national government. Nevertheless, this is justifiable since many government agencies would need that data; from the Department of Health (DOH), the Department of Social Welfare and Development (DSWD), the National Economic and Development Authority (NEDA), and the Office of the President in Malacañang (Gavilan, 2014). II. Sampling Methods Like what we have discussed earlier, sampling techniques fall under two (2) categories: probability sampling and non-probability sampling. Let us discuss further these two (2) categories (Baraceros, 2016). 1. Probability Sampling – this method relies on the researcher’s selection of respondents on pure chance thus, everybody in the population participates. a. Simple-random Sampling – the selection of respondents is based on pure chance. b. Systematic Sampling – this is a type of probability sampling scheme in which sample members from a larger population are chosen according to a random starting point and a fixed, periodic interval. This interval, called the sampling interval, is calculated by dividing the population size by the desired sample size. For example, a researcher wants to create a systematic sample of 1,000 students from a school with a student population of 10, 000. The researcher will then decide, a number, let us say like four (4), as the number to begin with. This means that the individual numbered “4” will his/her first selection, and then the researcher would decide that every tenth person from then on (desired sample size) would be included in the count. The sample would then be 4, 14, 24, 34, 44, and so on until s/he reaches the individual numbered “9,994”. c. Stratified Sampling – this type of method chooses a sample that later on be subdivided into- sub-groups during the stage of data analysis. For example, a researcher might divide a sample of people based on their religious affiliation, age, and others. This sampling technique is effective for studying how an issue might differ across subgroups. d. Cluster Sampling – this is choosing respondents in clusters, rather than in separate individuals. For example, selecting five (5) classes of 50 students each from a school that has a 50,000 students population. 2. Non-Probability Sampling – this technique is prone to bias due to the pre-selection of respondents. Below are the types of non-probability sampling techniques (Crossman, 2018). a. Quota Sampling – selecting specific samples that the researcher knows will correspond to the population in terms of one (1), two (2), or even more characteristics. b. Voluntary Sampling – choosing samples who are very much willing to participate as respondents. c. Purposive Sampling – selecting respondents who are deemed by the researcher as samples with good background and knowledgeable about the study being conducted. d. Availability Sampling – choosing respondents who are easy to find and willing to communicate with the researcher. e. Snowball Sampling – also known as chain sampling, is used when the respondents cannot be easily located like homeless individuals or street children. A snowball sample is one in 06 Handout 1 *Property of STI [email protected] Page 2 of 5  SH1806 which the researcher gathers data on the few members of the target population s/he can locate, then asks those personages to provide information needed to find other members of that population whom they know. III. Designing the Research Instruments Research instruments are measurement tools designed to acquire data and information on a topic or subject of interest from research subjects. Surveys, tests, questionnaires and the like are types of research instruments used to obtain data. The two (2) commonly used research instruments are surveys and interviews. Surveys usually are self-administered. The researchers allow respondents to answer predetermined sets of questions. On the other hand, interviews are done face-to-face or in any means that would allow the researcher and the respondent to converse and engage in a discussion on a particular topic. Interviews are often used in qualitative research primarily because it can be used to explore the views, experiences, beliefs, and motivations of the participants. We must remember that qualitative research involves the use on non-numerical data. In quantitative research, surveys are often used and are commonly set in such a way that it simulates a questionnaire. Surveys enable the researcher to collect numerical descriptions of trends, behaviors, attitudes, orientation, and opinion about a sample. Though most respondents feel uneasy when answering questions especially the personal ones, it is the job of the researcher to assure the respondents of confidentiality. There are setbacks as well since surveys could yield and produce biased results. Respondents may feel obliged to answer something that will please the researcher. Ambiguity is another issue of using surveys since the respondent may perceive the questions differently or may misunderstood the query. In order to curb and solve this issue, the researcher must formulate questions that can be easily understood and with less complicated terms. Guidelines in Formulating Surveys  Ethics should be observed. A few respondents might feel awkward answering the questions so it is the researchers’ responsibility to make sure that confidentiality will be exercised.  Consider the length of the survey. If the survey is long, respondents may feel an experience called “survey fatigue” which might cause them to answer haphazardly thus resulting in biased results.  Surveys must be free from ambiguity. Questions should be understandable. Avoid abbreviating terms as this could cause confusion in the part of the respondents.  Double-barreled questions such as those that tap several issues but require only one (1) answer. These questions may result in error in the behaviors being measured for the question. Here is an example of a double-barreled question: Do you think there should be more classes about Rizal and Philippine Constitution? The question is trying to answer two (2) different issues, which later on could negatively alter the results. Combining both issues in one (1) sentence or question makes it unclear what exactly is being measured. In order to reinforce the guidelines written above, qualtrics.com have made the following infographic about what makes a good survey question. 06 Handout 1 *Property of STI [email protected] Page 3 of 5  SH1806 Figure 1. What makes a good survey question Source: www.qualtrics.com Types of Survey Questions (Torneo & Clamor-Torneo, 2017) Questions included in surveys come in the following forms: 1. Open-ended questions – these questions do not have fixed answers or even options. Respondents are allowed to answer freely. In these types of questions, it is imperative that the researcher record the statements of the respondents in verbatim. The researcher must also carry out follow- up questions to clarify certain unclear and shady items and responses. Example: Why did you choose to elect the incumbent president? Kindly explain. 2. Dichotomous Questions – these questions may have two (2) possible answers. Often these questions have responses by either a yes or a no, true or false, or agree or disagree. These questions are employed when the researcher wants to categorize the respondents’ opinions or preferences. Example: Tattoos greatly affect a person’s skills at work. __ Agree __ Disagree 06 Handout 1 *Property of STI [email protected] Page 4 of 5  SH1806 3. Multiple-response Questions – these questions have multiple answers. Example: How were you able to know the programs being offered by STI College Ortigas-Cainta? __ Internet __ Radio advertisement __ Television commercial __ Word of mouth 4. Matrix Questions – these are questions which are open-ended and task the respondents to evaluate one (1) or several rows or choices. A Likert scale is a type of a matrix question. Example: Figure 2. Likert scale. Source: https://www.fieldboom.com/blog/likert-scale/ 5. Contingency Questions – these are questions intended for specific respondents only. Normally, these are follow-up questions provided that the respondent agrees to a certain question item. Example: Have you ever tried any illegal substances in the past? __ Yes __ No If yes, what illegal substances have you tried? __ Marijuana __ Crystal Meth __ Heroine __ Cocaine __ Ecstasy (tablet or liquid form) __ Others, please specify ________ References: Babbie, E. R. (2016). The basics of social research. Boston, MA: Cengage Learning. Baraceros, E. (2016). Practical research 2. Manila: Rex Bookstore Inc. Crossman, A. (2018, February 04). Find the best sampling technique for your research. Retrieved May 24, 2018, from https://www.thoughtco.com/sampling-designs-used-in-sociology-3026562 Gavilan, J. (2014, July 01). From numbers to action, thanks to the National Nutrition Survey. Retrieved May 24, 2018, from https://www.rappler.com/move-ph/issues/hunger/62031-national-nutrition-survey-role-fnri Sampling. (n.d.). Retrieved from https://www.merriam-webster.com/dictionary/sampling Clamor-Torneo, H. & Torneo, A. (2017). Practical research 2. Quezon City: Sibs Publishing House Inc. 06 Handout 1 *Property of STI [email protected] Page 5 of 5  SH1806 Looking for Answers in Data Collection I. Data Collection Data collection is the process of acquiring and gathering data about the variables the researcher/s are seeking to study. It is important in processing statistical output. Data gathering is essential in any field and discipline. As mentioned in the previous handouts, there are two (2) sources where we can collect data: primary and secondary data. Primary data These are first-hand data collected for a specific purpose. It can be collected through, interviews, questionnaire, experiments, or even observations. Interviews are an oral exchange of questions and answers. It may be done in focus groups or on a face-to-face basis. It can be flexible since the researcher could very well make follow-ups on answers the respondents have given. It can be costly for a researcher to conduct interviews but the data s/he will get is more accurate for his or her research. The most common way of collecting data is with questionnaires. It can be given personally or sent through e-mail and quite recently, administered using a social media platform. Questionnaires could be researcher-developed or adapted or adopted. Using the adapted or adopted, a researcher must first seek permission if s/he could use the ready-made questionnaire of if s/he can alter the questions so that it could be related to the study. If the researcher creates his/her own questionnaires, that research instrument should undergo validation and proofreading first before it could be administered. Secondary Data Data can also be acquired by reviewing existing documents made available to researchers. Secondary data may come from government agencies and international organizations. Compared with primary data, it is important that the researcher is familiar on how the secondary data was collected, what the response categories are for each question, who the population of the study was, and others. Probably the greatest advantage of using secondary data is economics. Someone else has collected the data which means the researcher does not have to devote time and release money when acquiring information. Also, since most of these data is in electronic form, the researcher can easily analyze the data and trim out the ones that do not have any relationship to his/her study. II. Data Organization In order for the data to be presented properly, it should be organized first. It may be organized and by encoding it on a spreadsheet. A researcher must make sure that the data he or she will encode should be complete and accurate. It is a good practice that if the data is incomplete or some of the values are missing, it should be verified first and leave it blank for the meantime. (Uy, Cabauatan, De Castro, & Grajo, 2016). 07 Handout 1 *Property of STI [email protected] Page 1 of 3  SH1806 Let us assume that you as a researcher who would want to organize a data you have acquired regarding the monthly rental of dormitories within the proximity of your STI school. See the table below which shows a data set for your reference. Dorm No. 1 Monthly rent in Peso P500.00 Proximity in meters 600 m Air-conditioned No Wi-Fi None CCTV None Dorm No. 2 Monthly rent in Peso P700.00 Proximity in meters 650 m Air-conditioned No Wi-Fi None CCTV Yes Dorm No. 3 Monthly rent in Peso P1500.00 Proximity in meters 200 m Air-conditioned Yes Wi-Fi Yes CCTV Yes Given the data sets above, we can now organize it. What we can do is assign the number one (1) to signify yes and zero (0) to signify no. This is done for ease in encoding. Dormitory Rent Proximity Air-conditioned Wi-Fi CCTV 1 P500.00 600 m 0 0 0 2 P700.00 650 m 0 0 1 3 P1500.00 200 m 1 1 1 III. Data Presentation The data that were collected will have to be summarized in preparation for its presentation. The presentation of data may be in the form of texts but usually, it is presented visually using graphs, tables, and charts. A table is a summary of data in columns. The simplest table is the two-column table where the first column shows the variables and the second one shows the occurrences. The frequency or distribution table may be employed by the researcher to show more variables or layers. The following are the examples of a two-column and frequency table. 07 Handout 1 *Property of STI [email protected] Page 2 of 3  SH1806 Table 1. A two-column table Source: http://blog.cardinalec.com/ Table 2. A Frequency table Source: http://psychstat3.missouristate.edu Charts and graphs may also be used to present data. This is normally done if you need to show data to an audience. The most common graphs are histogram, frequency polygon, line graph, and scatter diagram. Histogram and bar graphs and charts are used to for comparison of data. Line graphs, on the other hand, show the performance of a certain variable in a given period of time or occurrence. When pie charts and graphs are used, a circle is divided into sectors that each represent a proportion of the whole (Uy, et.al., 2016). References: Baraceros, E. (2016). Practical research 2. Manila: Rex Bookstore Inc. Clamor-Torneo, H. & Torneo, A. (2017). Practical research 2. Quezon City: Sibs Publishing House Inc. Uy, C., Cabauatan, R., De Castro, B., & Grajo, J. (2016). Practical research 2. Quezon City: Vibal Group, Inc. 07 Handout 1 *Property of STI [email protected] Page 3 of 3  SH1806 Looking for Answers in Data Collection I. Creating the Effective Survey Questions In creating survey questions, researchers must make sure that the questions are effective. There are multitudes of ways how we can make our questionnaires understandable to our respondents. It is important that the respondents understand each question so as to avoid any negative results. Pre-testing the survey Administer the survey you have formulated to a few potential respondents or even your friends and get feedback. Have your questionnaires validated and proofread by someone, like your teacher for grammar and content errors. Think About How You Will Administer your Survey Surveys can be administered via mail, phone, in-person, online, or through social media; and there are different questionnaire considerations for each mode. If the respondent is filling out a survey online or through social media, think about how it will look. Consider the following points: Give instructions – give specific instructions to the respondents if there are any. Make sure that your survey has clear instructions. Keep questionnaires short – respondents are less likely to answer long survey questionnaires. Avoid boredom. Keep question order in mind – survey responses can be impacted by previous questions. Meaning, as much as possible, each question should be related to each other. Have an introduction – if a respondent reads the survey, provide a title for each section. Start with general questions – it will warm up the respondents. Sensitive questions at the end – sensitive questions should best be reserved at the end of the questionnaire. Have them warm up first by answering general questions. Checklist for Formulating Effective Survey Questions (DeFranzo, 2011) Use direct and simple language. Use words that can easily be understood. Ask only one (1) question per item. If the question causes emotional discomfort, reword the question to soften the emotional response. General questions should be asked first. Sensitive ones go last. For questions about behavior or attitude, ensure all reasonable answers are included. Questions should be as specific as possible in relation to the research objectives. Make the question as specific as possible in terms of When, Where, and How. Ask questions of the respondents, as they will give more accurate responses when recalling information about themselves rather than for others. Reference: DeFranzo, S. (2011, October 17). Checklist for survey questionnaire design. Retrieved on July 18, 2018, from https://www.snapsurveys.com/blog/survey-questionnaire-design-checklist-for-factual-behavioral-questions/ 07 Handout 2 *Property of STI [email protected] 1 of 1  SH1806 The Math of Quantitative Research I. Why Use Statistics Statistics, as defined by University of North Carolina, is a scientific discipline that deals with the acquisition, organization, study, analysis, display, and interpretation of data. According to the university, statistics provides reasoning and methods for creating and understanding data. In the realm of quantitative research, there are two (2) forms of statistics used by researchers: the descriptive and inferential statistics. In inferential statistics, we try to determine if the hypothesis that we have formulated can be accepted or not. According to Clamor-Torneo and Torneo (2018), inferential statistics allows the researchers to draw reliable conclusions about a larger population based on a small sample, and allows the researcher to draw conclusions that go beyond the data collected. For this course, however, we will largely deal on descriptive statistics and its role on quantitative research. Descriptive statistics deals with summarizing and describing data. The goal of descriptive statistics is to calculate quantities and amounts that describe the collected data. It deals with the mean, median and mode to indicate the center of the data (Taylor, 2017). In addition, descriptive research aims to reduce the bulk of data collected into a more manageable, understandable, and simpler form without conceding the value of the data the researchers want to convey. For example, instead of trying to explain a multitude of data, a researcher might want to get the mean, median, or mode, as mentioned earlier, and present it using graphs, tables, and/or charts. It is often preferred by researchers since it permits them to easily describe basic patterns (Clamor-Torneo & Torneo, 2018). Descriptive statistics is important to qualitative research because when a study involves a long list of measures and figures, it chunks them down to understandable and explainable numbers. Also, as explained earlier, we can use graphs and charts to make it more appealing. To exemplify, look at the figure below. Figure 1. Incumbent and past Philippine presidents' net ratings Source: www.pressreader.com 08 Handout 1 *Property of STI [email protected] Page 1 of 3  SH1806 As seen in the graph, all the presidents’ ratings significantly dropped over time. It can be inferred that something drastic happened in those years were his/her rating suddenly plummeted—for instance, the case of former president Joseph Estrada with his impeachment and former president Gloria Macapagal-Arroyo’s “Hello Garci” controversy. Through descriptive research, we can easily target the trends and patterns in a given data. Reading and analyzing too much data can be exhausting to the research and the reader; hence, it must be presented in a manner that is pleasing to the eye. II. Statistical Treatment of Data The numbers that describe what average or typical is within a given data is called the measures of central tendency. There are three (3) measures, namely: mean, median, and mode. They are calculated differently (Argawal, 2009). Mean - Mean is the average value found in a sample or data set. It is often symbolized by the x- bar or x̄. The sample mean is useful because it permits the researcher to estimate what the whole population is doing, without having to survey everyone. The formula for mean is x̄ = (Σ xi ) / n where: x̄ stands for the “sample mean”, Σ means “add up”, xi “all of the x-values”, and n means “the number of items in the sample”. Example: A set of numbers is given as follows: 10, 15, 16, 17, 29, 32, and 40. First, add all the numbers. You will get 159. Next, count how many data sets there are. In this case, there are seven (7) sets. Since Σ is 159, and n is 7, the mean (x̄) will be 22.71. Median - Median is a number in statistics that tells us where the middle data set is. The median formula is {(n + 1) ÷ 2}th where: n is the number of items in the set, and th means the placement number. In order to get the median, arrange the data sets in an ascending order first. Then, find the middle number in the data sets. For example: 1, 2, 3, 4, 5, 6, 7. The median is obviously 4. We will get the same result if we use the formula. There are seven (7) data sets in the group. Therefore, n=7. {(7+1) ÷ 2}th = 4 or 4th data set. However, this only works if there are odd number of sets. But what about if the data sets have an even number of figures? 08 Handout 1 *Property of STI [email protected] Page 2 of 3  SH1806 For example, these are the given figures: 29, 34, 89, 10, 85, and 66. Remember, arrange the data sets in an ascending order first. Therefore, it will be 10, 29, 34, 66, 85, and 89. Then, locate the two (2) numbers in the middle, so those will be 34 and 66. Add the two (2) numbers, and divide it by 2. So, 34 + 66 = 100/2 = 50. But, wait, there’s more! If there is a large data set, like the one in the example, divide the number of items by 2, then subtract 1 to find the number that should be above and the number that should be below. Therefore 100/2 = 50 – 1 = 49. The middle two (2) numbers will have 49 items above and 49 below. Mode - The mode or modal value is the most common number in a data set. It occurs or appears often in a collected data. A data set can have no mode or several modes. None: 1, 2, 3, 4, 6, 8, 9. One (1) mode: Also known as unimodal: 1, 2, 3, 3, 4, 5. Two (2) modes: Also known as bimodal: 1, 1, 2, 3, 4, 4, 5. Three (3) modes: Also known as trimodal: 1, 1, 2, 3, 3, 4, 5, 5. More than one (1) = multimodal. So, if the data sets are 56, 57, 56, 58, 59, 90, 98, 98, 65, 45, 34, 34, 23, 23, 24, 33, 56, 67, 78, 87, 87, and 56, organize the numbers first in a way that the same sets will be grouped together. Then it will be like this: 23 23 24 33 34 34 45 56 56 56 56 57 58 59 65 67 78 87 87 90 98 98. The common number that appeared the most in the given data is 56. That is the mode. The ANOVA Test The ANOVA or the analysis of variance is a test to determine if the survey or the results of an experiment are significant. It helps the researcher in deciphering whether to reject the null hypothesis or accept the alternate hypothesis. There are two (2) ANOVA types, namely: the one- way or the two-way. One-way or two-way refers to the number of independent variables in an ANOVA test. Remember that one-way has one (1) independent variable with two (2) levels, and two-way has two (2) independent variables and can have multiple levels. For example, a one-way ANOVA can have one (1) independent variable which can be a brand of energy drink, while a two- way ANOVA can have two (2) independent variables which can be the brand of energy drink and its flavor. References: Agarwal, B. L. (2009). Basic statistics. New Delhi: New Age International Publishing Clamor-Torneo, H. & Torneo, A. (2017). Practical Research 2. Quezon City: Sibs Publishing House Inc. Taylor, C. (2017, October 18). A Beginner's Guide to Statistics. Retrieved July 25, 2018, from https://www.thoughtco.com/what-is-statistics-3126367 University of North Carolina, Department of Math and Statistics. (n.d.). What is statistics? Retrieved July 25, 2018, from https://www.uncg.edu/mat/undergraduate/whatisstatistics.html 08 Handout 1 *Property of STI [email protected] Page 3 of 3  SH1806 Conclusions and Recommendations I. Drawing Conclusions II. Formulating Recommendations In research, the conclusion is directly related to the main Making recommendations is making the readers know that research problem or objectives. It presents the contribution of there are real-life implications in the study a researcher has new learnings to the reader. In writing the conclusion, the chosen and making them know that there are ways how the researcher should state his/her conclusions with conviction. findings can be addressed. In writing the recommendations, Writing the conclusion also provides the researcher an avenue researchers must take into account the following: to persuade and restate the research problem. In the course of writing the body of the research, the researcher presents facts a. The recommendation should have the aim and effort to and, upon writing the conclusions, s/he tries to give the reader solve the problems of the study. the inferences about the topic. Some researchers are hesitant b. It should address the benefits it can give to society. to give out conclusions and will leave that task to the reader. c. The recommendations should be based on facts and Researchers should avoid that because they are the ones best evidences that were gathered. informed on aspects that influence the findings and outcomes Moreover, in this section, it is the researcher's opportunity to or their studies. show and discuss the actions that other and future researchers According to Amorado, Boholano, and Talili (2017), the should take as a result of the research. Always remember that following are important things researchers must consider in even if the results of the study are not favorable, other writing this part of the research: researchers can use the study as reference or as a guide. This 1. Summarize the main point you have tackled in the being said, recommendations can take two (2) forms, namely: introduction and review of related literature. recommendations for further study and recommendations for 2. Revisit the research methodologies and/or design you change; or even both. employed. 3. Discuss the implications of your findings. 4. Validate the theory presented. References: 5. Answer the problem. Amorado, R., Boholano, H., & Talili, I. (2017). Quantitative research: A practical approach. Malabon City, Philippines: The conclusion is the researcher's way of summing up all the Mutya Publishing House findings s/he unraveled. The conclusion allows the researcher Baraceros, E. (2016). Practical Research 2. Manila, Philippines: to remind the reader about the important concepts of the Rex Bookstore Inc. study. It also highlights key findings and provides a briefer of Clamor-Torneo, H. & Torneo, A. (2017). Practical research 2. what was discussed. The conclusion also allows the reader to Quezon City, Philippines: Sibs Publishing House Inc. see the "big picture" as it relates to the problem being solved. 09 Handout 1 *Property of STI [email protected] Page 1 of 1