Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
1.0 INTRODUCTION Architecture is primarily concerned with the development of the physical environment (built and natural) for optimum use by man. It is defined as the art and science building which is concerned with individual houses, large composite building complexes and even a whole city. Ultimately the art and science of organizing spaces for use by man. Research is simply the process of arriving at dependable solutions to problems through the planned and systematic collection, analysis, and interpretation of knowledge of his environment, to accomplish his purpose and to resolve his conflicts. Architectural research in Nigeria is geared towards improving the quality of architectural products in the country and in achieving there is a need to focus on areas having such defects in qualitative architecture within the country. The sampling methods are used for proper navigation and focus on areas needing improvement. Sampling is the process of selecting units (e.g., people, organizations) from a population of interest so that by studying the sample we may fairly generalize our results back to the population from which they were chosen. There is no concept as fundamental to the conduct of research and the interpretation of its results as in sampling. Except when a complete census is taken, research is almost invariably conducted by means of a sample, o the basis of which generalizations applicable to the population from which the sample was obtained are reached. (Osuala 1987). In research work, sampling has become very important because of the extreme difficulty involved in listing, observing and reaching every element in the population about which every opinion is to be expressed. It is the process of selecting representative elements (samples) from a given population. The process is such that it enables an investigator to choose elements which in number and character sufficiently reflects the relevant features of the population from which they are drawn. The purpose of this process is to provide a realistic basis upon which generalizations about the population may be drawn from sample characteristics. (Olu Ojo 2003). 1 1.1 BASIC TERMINOLOGIES IN SAMPLING (a) Population: This consists of all conceivable elements, subjects and observations that are of primary interest to a researcher or a study. (Olu Ojo 2003) (b) Study Population: The aggregation of elements from which the sample is actually selected. (Olu Ojo 2003) (c) Sample: This consists of selected elements or observations from a population for the purpose of study with the hope that the sample is representing the population. (Olu Ojo 2003) (d) Subject / Element: The individual member of a population or sample. (Olu Ojo 2003) (e) Sampling Unit: The element or set f elements considered for selection from a sampling population. (Olu Ojo 2003) (f) Sampling Frame: The actual list of sampling units from which the sample, or some stage of the sample selected. (Olu Ojo 2003) (g) Observation Unit: This is an element of aggregation of elements from which information is collected. (Olu Ojo 2003) (h) Variable: This is a set of mutually exclusive attributes ; sex, age, religion, race, etc. (Olu Ojo 2003) (i) Parameter: The summary description of a given variable in a population. (Olu Ojo 2003) (j) Statistics: Summary description of a given variable in a sample. (Olu Ojo 2003). 2 2.0 SAMPLING METHODS Sampling methods involves a number of activities which enhances and increases the effectiveness of the entire process. These activities include: (i) Problem Definition: Sampling is undertaken with a view to choosing from a population a number of elements, which reflects the relevant features of the population. It is thus only logical that the first test in any sampling activity is to define the variables that are crucial to the kind of information about the population. These variables reflect characteristics that will enable the investigator to obtain data at all aspects of the desired information. (Olu Ojo 2003) (ii) Data Collection: Having defined the problem, it is important to proceed to actual sample selection on the basis of determined characteristics. Sample selection in this regards follow from whatever information that is available about the population, research variables and the number of elements to be drawn to represent the population. Then relevant observation is made of the sample drawn based on the research variables. From the observations made, analysis is made to obtain relevant information regarding trends and relationship and making generalizations and conclusion there upon. It is also possible at this stage to ascertain true analysis whether or not sample drawn and data gathered are sufficient, reliable and relevant to the problem at hand. (iii) Generalization: Since the ultimate aim of sampling is to provide reasonable basis for the drawing of conclusion about a population from the samples collected from it, generalization becomes a determining point in sampling undertaken. Although it does not form part of sampling, it holds a critical implication in that it is the ultimate basis for the evaluation of the representativeness, completeness and reliability of the sample. 3 2.1 PRINCIPLES OF SAMPLING Sampling is governed by two basic laws which ensures their representativeness. These laws include: (i) Principle of Statistical Regularity: This principle states that the more randomly a sample is drawn from a population, the more the likelihood the sample reflects the characteristic of that population. In other words, it cautions that for sample to truly reflect population characteristics, such must be selected through a process that guarantees each element the probability of being included in the sample. (ii) The Principle of Inertia of Large Number: It holds that the larger the number of elements (samples) drawn to represent a population the greater the likelihood of that sample reflecting the true features of that population. 2.2 TYPES OF SAMPLING METHODS Sampling methods generally describe a particular means of drawing samples from a given population, that is, techniques and procedures adopted to provide a basis for generalizing about a population through a sample taken from it. Typically divided into probability an nonprobability techniques. (Olu Ojo 2003). 2.2.1 PROBABILITY SAMPLING METHOD These are those designed to guarantee every element of a population an equal and independent chance of being included in the sample drawn. These methods are generally referred to as unbiased and objective considering them against the against the non-probabilistic methods. That is the researcher has preference of selection. (Olu Ojo 2003). There are four basic types f probability sampling methods and these include: simple random sampling, systematic sampling, stratified sampling and cluster/ area sampling. 4 (1) SIMPLE RANDOM SAMPLING: This is the basic method on which all other probability sampling methods are built and they all involve it at some stage of the process. This requires that each element of the population have an equal chance of being selected. A simple random sample is one selected not on the basis of any criteria, or set of criteria but o the basis that the researcher closes his eyes and mind but randomly picks elements that constitute the sample. It is a method in which samples are drawn from a comprehensive list of population elements without any pre-specified order capable of destroying the balance and opportunity of including every element in the population. This is done through one of the following methods: (a) Balloting or open basket: This is to identify all the elements, give each element a name or number, writing such name/number on a small piece of paper. Squeezing the paper, picking is done one after the other until the required sample number is gotten. (b) By giving numbers or names to all the elements and writing the numbers or names on cards shuffling them at intervals and each time the cards are shuffled picking the card on top until the required number for the sample is gotten. (c) The use of table of random numbers or digits. A table of random numbers specially designed for sampling purposes is found in most books on statistics. By this method, all elements of the population should be numbered. Digits are selected from the table in any systematic way and these elements whose numbers coincide with random digits are included in the sample. This process is continued until the required sample number is gotten. Advantages of Simple Random Sampling (i) This procedure ensures that every sampling unit of the population has an equal chance of being included in the sample. (ii) Makes easy in analyzing data and computing statistics. Disadvantages of Simple Random Sampling (i) Requires the researcher to have knowledge of population and listing it. 5 (ii) Requires locating and indentifying each chosen element for study. This can consume much of researcher’s time, money and other resources. (2) SYSTEMATIC SAMPLING This is an extension of the simple random sampling method. Systematic random sampling involves the selection of every Nth item from serially listed population subject or units after the first sampling unit is selected at random from the total of sampling units. Nth item refers to elements of the population that are lying at each interval point in the list of elements. Systematic sampling proceeds to choose the sample by first determining the sample interval, that is, the standard distance between elements selected I the sample, which is obtained by dividing the study population by the sampling size. (Olu Ojo 2003). Advantages of Systematic Sampling (i) It is more convenient than simple random sampling because it makes sample very simple to draw. (ii) It is also more amenable for use with very large populations or when large samples are to be selected. (iii) If population list is ordered with a variable related t what is being studied, systematic sampling has the effect of stratification on that variable. (iv) Another advantage is that in an alphabetized list of names, this approach will avoid repeating sampling from the same family. Disadvantage of Systematic Sampling (i) It may result to non-representativeness if the Nth is related to a periodic ordering in the population listing. (ii) System sampling is not completely random once a random starting point has been selected, all other elements are pre-determined. 6 (iii) Neglecting to take stratification effect into account in the statistics, systematic sampling will yield estimates of accuracy that are too low. (3) STRATIFIED SAMPLING This is an applied random sampling method. It is a modification of simple random and systematic sampling techniques. It attempts to overcome the problems of these techniques by categorizing elements of the population into strata on the basis of vital characteristics likely to affect the sampling design. By this method, the researcher identifies the key characteristics which will be of importance to the research. The population is then divided into groups along these characteristics and each group is known as a stratum. Thereafter, it proceeds to select samples for each stratum in proportion to sample’s share of the total population based on the simple random or systematic criteria. This is known as proportionate stratified sample because the sample size drawn from each sample is proportional to the population size of the stratum. (Olu Ojo 2003). Advantages of Stratified Sampling (i) It ensures representativeness of whatever characteristics is used to classify units into strata. (ii) Facilitates analysis of strata and hence of subgroups of population; and eases comparison of subgroups. (iii) It is practically convenient and considerably reduces considerably reduces the cost of execution. (iv) The overall picture or the conclusion drawn for the whole population is likely to be more accurate than if we had simply taken a sample without stratification. Disadvantages of Stratified Sampling (i) It requires accurate information on proportion of population in each stratum, or else error is increased. 7 (ii) If information for stratification is not available, it may make it costly to obtain and prepare lists. (iii) Risks improper classification of individual in strata due to clerical error or poor measurement. (iv) An extensive sampling frame is necessary and those strata levels of importance can only be selected subjectively. (4) CLUSTER OR AREA SAMPLING This is a sampling method which can be employed where no sampling frame exists, and often for a population which is distributed over some geographical areas. I t employs principles similar to those of stratification. However, rather than apply these principles on the basis of feature of population elements, it applies them to the study region by dividing it into blocks and clusters for selection purposes. From each cluster is selected sample to reflect key features of cluster. This is therefore appropriate where research is conducted in a large territory across which samples are to be selected. The technique involves selecting one or more geographical areas and sampling all the members of the target population that can be identified. In other words, cluster sampling is a sampling procedure in which the units of analysis in the population are grouped into clusters and a sample of cluster serves as a basis for sampling rather than a sample of individual. Advantages of Cluster Sampling (i) It reduces travel if clusters are used for interviews. (ii) It reduces construction of sampling frames. (iii) It permits studies of individual clusters and comparison of clusters. (iv) It can use other persons for follow-up within a cluster if cluster was sampled, or can use other clusters. 8 Disadvantages of Cluster Sampling (i) It may not be efficient. The closeness of clusters may affect representativeness. (ii) It may result in larger errors in estimating population values than other probability sampling methods. (iii) It requires that each member of population be assigned uniquely to cluster, otherwise it may omit or duplicate cases. 2.2.2 NON-PROBABILITY SAMPLING METHODS The distinguishing characteristics of non probability sampling techniques are the absence of each element being included in the sample selected. The sampling technique therefore adopted rests squarely on subjective criteria determined by particular purpose served or specific circumstance in which the sample has to be selected. Although sometimes, all the elements end up having equal opportunity. This may be the result of accident and not by design because the population elements are not deliberately given equal chance of being selected. (Olu Ojo 2003). Four of the non-probability sampling techniques commonly used shall be examined in this write-up and they include; purposive or judgmental sampling, accidental or convenience sampling, quota sampling and snowball sampling. (1) PURPOSIVE OR JUDGEMENTAL SAMPLING: This is a non-probability sampling technique in which the selection of sample elements from a given population is based entirely on the subjective choice of the researcher. The sample selection is therefore the result of the researcher’s opinion as to which element best provide desired basis and probability of good outcome. Experience, wisdom, research purpose and solicited opinions are usually relied upon on sample selection by purposive technique. This technique is also referred to as judgment or judgmental sampling. (Olu Ojo 2003). 9 (2) ACCIDENTAL OR COVENIENCE SAMPLING: This is the type of sampling adopted whereby elements are drawn into sample on first come first serve basis. This is usually employed where the population size is unknown and is impossible or difficult to determine with accuracy. Consequently, elements are selected as they are observed. In other words, sample elements are selected on the bass of ease of data collection, data analysis or both. That is, rather research purpose or considered opinions regarding representativeness, it is the ease of procedure and convenience of the researcher that forms the criteria for selection. This type of sampling is most useful when a researcher simply wants to have a smattering idea about a phenomena and not anything accurate. Accidental or convenience sampling is very cheap and does not consume much time. However it is most time unrealistic and reliable and lacks precision. There caution should be exercised in using accidental or convenience sampling. (Olu Ojo 2003). (3) QUOTA SAMPLING: Quota is to non-probability sampling method what stratified sampling is to probability sampling method. All he procedures are similar for both sampling and stratified methods. The only difference is that in quota sampling after the stratification is made, a non random method is applied in selecting the elements of the sample. In quota sampling, population elements are drawn into sample in specific sub-sections representing vital population characteristics or features relevant to the problem at hand.(Olu Ojo 2003).. This is obtained in two stages: (a) A matrix of the population is drawn based on the specified characteristics or features to be reflected in the sample selection. (b) From each of the matrices, a sample is drawn that is proportionate to the size of that section relative to the population. The application of the quota sampling gives, semblance of representativeness though it is a non-probability sampling method. Quota sampling has many inherent problems. 10 First, the quota frame (the proportion that different cells represent) must be accurate and it is often difficult to get up-to-date information for this purpose. Second, biases may occur in the selection of ample elements with a given cell-even though its proportion in the population is accurately estimated. (4) SNOWBALL SAMPLING: The method in snowball sampling is to reach is to reach other sample elements through initial elements previously included in the sample. Snowball sampling is a method through which the researcher develops an ever-increasing set of observations. The researcher asks one participant in the event under study to recommend others for interviewing, and each of the subsequently interviewed participants is asked for further recommendations. For instance, if you wish to learn the patter of recruitment into a religious organization over time, you might begin y interviewing fairly recent converts, asking them who introduced them into the group. You might the interview the persons named, asking them who introduced them into the group. You might then interview the persons named asking, in part, who introduced them and so forth. In this case your sample would “snowball” as each of your interviewees suggests others. (Olu Ojo 2003). 2.3.3 WHEN TO USE NON-PROBABILITY SAMPLING TECHNIQUES In general, probability sampling is more preferable to non-probability sampling because the sampling error is smaller in probability sampling. In practice, there are researches or sampling situations, where only non-probability sampling methods can be used especially, where there are enough justification for it or where probability sampling method is not feasible. Situations where non-probability sampling methods could be used include: (a) When it is impossible to reach some elements within the population or inexact and unknown population where the population elements can only be imagined. (b) When there is a possibility that a probability sampling may not give a fair representation of a population. 11 (c) Where generalization of results is not necessary or may not be intended. In such a case the researcher may not be concerned whether or not the sample is a representative of the population. (d) Cost and time are required to conduct the study using probability sampling are much more than that required by non-probability sampling. Enough money and time may not be available for the researcher. 12 3.0 OTHER SAMPLING METHODS In addition to the sampling methods discussed so far, there are other sampling methods that can be useful in certain circumstances with any of the probability or non-probability sampling methods aforementioned. These include: multistage sampling, sequential sampling, double sampling, key informant sampling and representation sampling. (i) Multistage Sampling: This is an extension of cluster sampling which involves selecting samples in two or more stages. Multistage sampling method can be employed in a situation where the researcher finds out that the population is so complex that he needs more than one sampling technique to select his sample. He definitely has to result to sampling in stages. This method requires the researcher to choose his sample in stages until he gets his required sample. (Olu Ojo 2003). (ii) Sequential Sampling: Here the sample size is not fixed in advance. Sequential sampling allow one to start with a small sample and then continue sampling until a criterion of adequacy is met. Typically, in sequential sampling, an initial sample is taken, and the data are analyzed to see if the needed statistical precision has been obtained or whether a larger sample is needed. If the latter, we obtain additional cases until the desired precision is reached. (Olu Ojo 2003). (iii) Double Sampling: The use of double sampling can be used to obtain a more representative sample in situations where sizeable non-respondent group exist. This increases the likelihood of obtaining a more representative sample. Double sampling permits a check on reliability of information obtained from the sample survey of a large sample, another sample can be drawn from the same group for a more comprehensive investigation through sampling intensity. (Olu Ojo 2003). (iv) Key Informant Sampling: 13 Here key informants are selected from the total population as part of the sample. The elements so selected must possess certain features or characteristics which others do not have. For instance they must possess knowledge, skill and willingness to give out information. (Olu Ojo 2003). (v) Representation Sampling: This is a situation where each element in a sample represents an identified group. For example in a cooperative society, a researcher an talk to the president, secretary, treasurer and two other members of the cooperative as representatives of the group. (Olu Ojo 2003). 14 4.0 CONCLUSION In conclusion architectural research in Nigeria carried out by researchers employ one or more of the above discussed sampling methods in other to focus properly on target population, geographical location and other related architectural problems so that at the end research is done more effectively, saving time and cost. 15 REFERENCES Ogunsote O.O (2009) Lecture note on research methods in architecture, F U T Akure, Ondo state Olu .O (2003) Fundamentals of research methods, Nelson Clammy press, Ibadan, Oyo state. Osuala E.C (1987) Introduction to research methodology, Africana-Fep publishers, Uyo, Akwa ibom state. 16