DOCSLIB.ORG

The Importance of Replication in Wildlife Research

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
The Importance of Replication in Wildlife Research University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln USGS Northern Prairie Wildlife Research Center US Geological Survey 2002 The Importance of Replication in Wildlife Research Douglas H. Johnson USGS Northern Prairie Wildlife Research Center, [email protected] Follow this and additional works at: https://digitalcommons.unl.edu/usgsnpwrc Part of the Other International and Area Studies Commons Johnson, Douglas H., "The Importance of Replication in Wildlife Research" (2002). USGS Northern Prairie Wildlife Research Center. 228. https://digitalcommons.unl.edu/usgsnpwrc/228 This Article is brought to you for free and open access by the US Geological Survey at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in USGS Northern Prairie Wildlife Research Center by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. InvitedPaper: THE IMPORTANCEOF REPLICATIONIN WILDLIFE RESEARCH DOUGLASH. JOHNSON,1U.S. GeologicalSurvey, Northern Prairie Wildlife Research Center,Jamestown, ND 58401, USA Abstract:Wildlife ecology and management studies have been widely criticized for deficiencies in design or analy- sis. Manipulative experiments-with controls, randomization, and replication in space and time-provide power- ful ways of learning about natural systems and establishing causal relationships, but such studies are rare in our field. Observational studies and sample surveys are more common; they also require appropriate design and analy- sis. More important than the design and analysis of individual studies is metareplication: replication of entire stud- ies. Similar conclusions obtained from studies of the same phenomenon conducted under widely differing condi- tions will give us greater confidence in the generality of those findings than would any single study, however well designed and executed. JOURNALOF WILDLIFEMANAGEMENT 66(4):919-932 Key words: control, experiment, metareplication, observational study, pseudoreplication, randomization, replica- tion, sample survey, science. Wildlife researchers seem to be doing everything assigned to attack, others to defend those articles. wrong. Few of our studies employ the hypotheti- We identified substantial problems in the design, co-deductive approach (Romesburg 1981) or gain analysis, or interpretation in nearly all of those the benefits from strong inference (Platt 1964). influential and highly regarded studies. We continually conduct descriptive studies, rather Despite all our transgressions, we must be than the more effective manipulative studies. We doing something right. We have brought some rarely select study areas at random, and even less species back from the brink of extinction. The often do the animals we study constitute a random bald eagle (Haliaeetus leucocephalus),whooping sample. We continue to commit pseudoreplication crane (Grus americana), Aleutian Canada goose errors (Hurlbert 1984, Heffner et al. 1996). We (Branta canadensis leucopareia), and gray wolf confuse correlation with causation (Eberhardt (Canis lupus) were extremely rare over much or 1970). Frequently we measure the wrong vari- all of their range only a few years ago; now they ables such as indices to things we really care are much more common. Many of us had given about (Anderson 2001). And we may measure up on the black-footed ferret (Mustela nigripes) them in the wrong places (convenience sampling; and California condor (Gymnogypscalifornianus), Anderson 2001). We often apply meaningless species that, while still at risk, appear to be recov- multivariate methods to the results of our studies ering. And we can manage for abundance if we (Rexstad et al. 1988). We test null hypotheses that want to, such as we have done for white-tailed not only are silly but are known to be false (Cher- deer (Odocoileusvirginianus) and mallards (Anas ry 1998,Johnson 1999, Anderson et al. 2000). We platyrhynchos).Recently, Jack Ward Thomas spoke rely on nonparametric methods that are neither of the "tremendous record of success" in our necessary nor appropriate (Johnson 1995). field (Thomas 2000:1). Such problems permeate our field. In my early Why this apparent inconsistency between our years as a hypercritical statistician, I read many error-prone methods and the successes of our articles in TheJournal of WildlifeManagement and profession? I hope to address that question here related journals. In virtually every article, I found by discussing what truly is important in scientific problems-often serious ones-in the methods research. I first discuss causation, then manipula- used to analyze data. That experience was repeat- tive experimentation as a powerful way of learn- ed later in a class in evolutionary ecology. During ing about causal mechanisms. The 3 cornerstones that class, we critically reviewed many key papers of experimentation are control, randomization, in evolutionary ecology. Some students were and replication. These features also are integral to observational studies and sample surveys, which are more common in our field. For those I E-mail:[email protected] types of studies especially, I argue that the most 919 920 IMPORTANCEOF REPLICATION* Johnson J. Wildl. Manage. 66(4):2002 important feature is replication. Further, I ex- where Yt(u) is the number of squirrels in woodlot pand this concept to the level of metareplica- u after the treatment, and Y (u) is the number of tion-replication of entire studies-and suggest squirrels in that woodlot if the treatment had not that this is the most reliable method of learning been applied (I follow Rubin [1974] and Holland about the world. It is a natural way of human [1986] here). If the woodlot is logged, then you thinking and is consistent with a Bayesian can observe Y( u) but not Y( u). If the treatment approach to statistical inference. Metareplication is not applied, then you can observe Y( u) but not allows us to exploit the values of small studies, Yt(u). Thus arises the fundamental problem of each of which individually may be unable to causal inference: you cannot observe the values reach definitive conclusions. Metareplication of Yt(u) and Yc(u) on the same unit. That is, any provides us greater confidence that certain rela- particular woodlot is either logged or not. tionships are general and not specific to the cir- Holland (1986) described 2 solutions to this cumstances that prevailed during a single study. problem. With the first, one has 2 units (u1 and u2, here woodlots) and assumes they are identical. CAUTION ABOUT CAUSATION Then the treatment effect Tis estimated to be The "management" in "wildlife management" implies causality.We believe we can perform some T= Y,(ul)- Y(2), (2) management action that will produce a pre- dictable response by wildlife. Even if the causes where u1 is treated and u2 is not. This approach is cannot be manipulated, it is useful to know the based on the very strong assumption that the 2 mechanisms that determine certain outcomes, woodlots, if not logged, would have the same such as that spring migration of birds is a response number of squirrels, that is, Y( u2) = Yc(ul). That to increasing day length, or that drought reduces assumption is not testable, of course, because 1 the number of wetland basins that contain water. woodlot had been logged. It can be made more The concept of causation is most readily adopt- plausible by matching the 2 units as closely as pos- ed in the physical sciences, where models of the sible or by believing that the units are identical. behavior of atoms, planets, and other inanimate That latter belief comes more easily to physicists objects are applicable over a wide range of con- thinking about molecules than to ecologists ditions (Barnard 1982) and the controlling factors thinking about woodlots, however. are few (e.g., pressure and temperature are suffi- Holland (1986) termed the other solution sta- cient to determine the volume of a gas). In the tistical. One gets an expected, or average, causal physical sciences, causality implies lawlike neces- effect T over the units in some population: sity. In many fields, however, notions of causality reduce to those of probability, which suggests T= E(Y,- Y), (3) exceptions and lack of regularity. Here, causation means that an action "tends to make the conse- where, unlike with the other solution, different quence more likely, not absolutely certain" (Pearl units can be observed. The statistical solution 2000:1). This is so in wildlife ecology because of replaces the causal effect of the treatment on a the multitude of factors that influence a system. specific unit, which is impossible to observe, by For example, liberalizing hunting regulations for a the averagecausal effect in the population, which species tends to increase harvest by hunters. In any is possible to estimate. specific instance, liberalization may not result in This discussion reflects the need for a control, an increased harvest because of other influences something to compare with the treated unit, such as population size of the species, weather which is required for either approach. To follow conditions during the hunting season, and the the statistical approach, we often invoke random- cost of gasoline as it affects hunter activity. ization. If, for example, we are to compare squir- Suppose you want to determine the effect on rel numbers on a treated woodlot and an squirrel abundance of some treatment (= puta- untreated one, we might get led astray if the tive cause), for example, selective logging in a woodlots were of very different size, or if one con- woodlot by removing all trees greater than 45 cm tained more mast trees, or if one was rife with diameter at breast height (dbh). The treatment predators of squirrels and the other was not. One effect on some woodlot can be defined as way-but not the only way-to protect against this possibly misleading outcome is to determine T= Y,(u) - Y(u), at random which woodlot receives the treatment J. Wildl. Manage. 66(4):2002 IMPORTANCEOF REPLICATION* Johnson 921 and which does not.
Load more

Recommended publications
  • Understanding Replication of Experiments in Software Engineering: a Classification Omar S
    Understanding replication of experiments in software engineering: A classification Omar S. Gómez a, , Natalia Juristo b,c, Sira Vegas b a Facultad de Matemáticas, Universidad Autónoma de Yucatán, 97119 Mérida, Yucatán, Mexico Facultad de Informática, Universidad Politécnica de Madrid, 28660 Boadilla del Monte, Madrid, Spain c Department of Information Processing Science, University of Oulu, Oulu, Finland abstract Context: Replication plays an important role in experimental disciplines. There are still many uncertain-ties about how to proceed with replications of SE experiments. Should replicators reuse the baseline experiment materials? How much liaison should there be among the original and replicating experiment-ers, if any? What elements of the experimental configuration can be changed for the experiment to be considered a replication rather than a new experiment? Objective: To improve our understanding of SE experiment replication, in this work we propose a classi-fication which is intend to provide experimenters with guidance about what types of replication they can perform. Method: The research approach followed is structured according to the following activities: (1) a litera-ture review of experiment replication in SE and in other disciplines, (2) identification of typical elements that compose an experimental configuration, (3) identification of different replications purposes and (4) development of a classification of experiment replications for SE. Results: We propose a classification of replications which provides experimenters in SE with guidance about what changes can they make in a replication and, based on these, what verification purposes such a replication can serve. The proposed classification helped to accommodate opposing views within a broader framework, it is capable of accounting for less similar replications to more similar ones regarding the baseline experiment.
    [Show full text]
  • Comparison of Response Surface Designs in a Spherical Region
    World Academy of Science, Engineering and Technology International Journal of Mathematical and Computational Sciences Vol:6, No:5, 2012 Comparison of Response Surface Designs in a Spherical Region Boonorm Chomtee, John J. Borkowski Abstract—The objective of the research is to study and compare Where X is the design matrix, p is the number of model response surface designs: Central composite designs (CCD), Box- parameters, N is the design size, σ 2 is the maximum of Behnken designs (BBD), Small composite designs (SCD), Hybrid max designs, and Uniform shell designs (USD) over sets of reduced models f′′(x)( X X )−1 f (x) approximated over the set of candidate when the design is in a spherical region for 3 and 4 design variables. points. The values of the two criteria were calculated using The two optimality criteria ( D and G ) are considered which larger Matlab software [1]. values imply a better design. The comparison of design optimality criteria of the response surface designs across the full second order B. Reduced Models model and sets of reduced models for 3 and 4 factors based on the The set of reduced models is consistent with the definition two criteria are presented. of weak heredity given in Chipman [2]. That is, (i) a quadratic 2 Keywords—design optimality criteria, reduced models, response xi term is in the model only if the xi term is also in the model surface design, spherical design region and (ii) an interaction xi x j term is in the model only if the xi or x or both terms are also in the model.
    [Show full text]
  • Relationships Between Sleep Duration and Adolescent Depression: a Conceptual Replication
    HHS Public Access Author manuscript Author ManuscriptAuthor Manuscript Author Sleep Health Manuscript Author . Author manuscript; Manuscript Author available in PMC 2020 February 25. Published in final edited form as: Sleep Health. 2019 April ; 5(2): 175–179. doi:10.1016/j.sleh.2018.12.003. Relationships between sleep duration and adolescent depression: a conceptual replication AT Berger1, KL Wahlstrom2, R Widome1 1Division of Epidemiology and Community Health, University of Minnesota School of Public Health, 1300 South 2nd St, Suite #300, Minneapolis, MN 55454 2Department of Organizational Leadership, Policy and Development, 210D Burton Hall, College of Education and Human Development, University of Minnesota, Minneapolis, MN 55455 Abstract Objective—Given the growing concern about research reproducibility, we conceptually replicated a previous analysis of the relationships between adolescent sleep and mental well-being, using a new dataset. Methods—We conceptually reproduced an earlier analysis (Sleep Health, June 2017) using baseline data from the START Study. START is a longitudinal research study designed to evaluate a natural experiment in delaying high school start times, examining the impact of sleep duration on weight change in adolescents. In both START and the previous study, school day bedtime, wake up time, and answers to a six-item depression subscale were self-reported using a survey administered during the school day. Logistic regression models were used to compute the association and 95% confidence intervals (CI) between the sleep variables (sleep duration, wake- up time and bed time) and a range of outcomes. Results—In both analyses, greater sleep duration was associated with lower odds (P < 0.0001) of all six indicators of depressive mood.
    [Show full text]
  • Topics in Experimental Design
    Ronald Christensen Professor of Statistics Department of Mathematics and Statistics University of New Mexico Copyright c 2016 Topics in Experimental Design Springer Preface This (non)book assumes that the reader is familiar with the basic ideas of experi- mental design and with linear models. I think most of the applied material should be accessible to people with MS courses in regression and ANOVA but I had no hesi- tation in using linear model theory, if I needed it, to discuss more technical aspects. Over the last several years I’ve been working on revisions to my books Analysis of Variance, Design, and Regression (ANREG); Plane Answers to Complex Ques- tions: The Theory of Linear Models (PA); and Advanced Linear Modeling (ALM). In each of these books there was material that no longer seemed sufficiently relevant to the themes of the book for me to retain. Most of that material related to Exper- imental Design. Additionally, due to Kempthorne’s (1952) book, many years ago I figured out p f designs and wrote a chapter explaining them for the very first edition of ALM, but it was never included in any of my books. I like all of this material and think it is worthwhile, so I have accumulated it here. (I’m not actually all that wild about the recovery of interblock information in BIBs.) A few years ago my friend Chris Nachtsheim came to Albuquerque and gave a wonderful talk on Definitive Screening Designs. Chapter 5 was inspired by that talk along with my longstanding desire to figure out what was going on with Placett- Burman designs.
    [Show full text]
  • Lecture 19: Causal Association (Kanchanaraksa)
    This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike License. Your use of this material constitutes acceptance of that license and the conditions of use of materials on this site. Copyright 2008, The Johns Hopkins University and Sukon Kanchanaraksa. All rights reserved. Use of these materials permitted only in accordance with license rights granted. Materials provided “AS IS”; no representations or warranties provided. User assumes all responsibility for use, and all liability related thereto, and must independently review all materials for accuracy and efficacy. May contain materials owned by others. User is responsible for obtaining permissions for use from third parties as needed. Causal Association Sukon Kanchanaraksa, PhD Johns Hopkins University From Association to Causation The following conditions have been met: − The study has an adequate sample size − The study is free of bias Adjustment for possible confounders has been done − There is an association between exposure of interest and the disease outcome Is the association causal? 3 Henle-Koch's Postulates (1884 and 1890) To establish a causal relationship between a parasite and a disease, all four must be fulfilled: 1. The organism must be found in all animals suffering from the disease—but not in healthy animals 2. The organism must be isolated from a diseased animal and grown in pure culture 3. The cultured organism should cause disease when introduced into a healthy animals 4. The organism must be re-isolated from the experimentally infected animal — Wikipedia Picture is from http://en.wikipedia.org/wiki/Robert_Koch 4 1964 Surgeon General’s Report 5 What Is a Cause? “The characterization of the assessment called for a specific term.
    [Show full text]
  • Blinding Analyses in Experimental Psychology
    Flexible Yet Fair: Blinding Analyses in Experimental Psychology Gilles Dutilh1, Alexandra Sarafoglou2, and Eric–Jan Wagenmakers2 1 University of Basel Hospital 2 University of Amsterdam Correspondence concerning this article should be addressed to: Eric-Jan Wagenmakers University of Amsterdam Department of Psychology, Psychological Methods Group Nieuwe Achtergracht 129B 1018XE Amsterdam, The Netherlands E–mail may be sent to [email protected]. Abstract The replicability of findings in experimental psychology can be im- proved by distinguishing sharply between hypothesis-generating research and hypothesis-testing research. This distinction can be achieved by preregistra- tion, a method that has recently attracted widespread attention. Although preregistration is fair in the sense that it inoculates researchers against hind- sight bias and confirmation bias, preregistration does not allow researchers to analyze the data flexibly without the analysis being demoted to exploratory. To alleviate this concern we discuss how researchers may conduct blinded analyses (MacCoun & Perlmutter, 2015). As with preregistration, blinded analyses break the feedback loop between the analysis plan and analysis out- come, thereby preventing cherry-picking and significance seeking. However, blinded analyses retain the flexibility to account for unexpected peculiarities in the data. We discuss different methods of blinding, offer recommenda- tions for blinding of popular experimental designs, and introduce the design for an online blinding protocol. This research was supported by a talent grant from the Netherlands Organisation for Scientific Research (NWO) to AS (406-17-568). BLINDING IN EXPERIMENTAL PSYCHOLOGY 2 When extensive series of observations have to be made, as in astronomical, meteorological, or magnetical observatories, trigonometrical surveys, and extensive chemical or physical researches, it is an advantage that the numerical work should be executed by assistants who are not interested in, and are perhaps unaware of, the expected results.
    [Show full text]
  • Section 4.3 Good and Poor Ways to Experiment
    Chapter 4 Gathering Data Section 4.3 Good and Poor Ways to Experiment Copyright © 2013, 2009, and 2007, Pearson Education, Inc. Elements of an Experiment Experimental units: The subjects of an experiment; the entities that we measure in an experiment. Treatment: A specific experimental condition imposed on the subjects of the study; the treatments correspond to assigned values of the explanatory variable. Explanatory variable: Defines the groups to be compared with respect to values on the response variable. Response variable: The outcome measured on the subjects to reveal the effect of the treatment(s). 3 Copyright © 2013, 2009, and 2007, Pearson Education, Inc. Experiments ▪ An experiment deliberately imposes treatments on the experimental units in order to observe their responses. ▪ The goal of an experiment is to compare the effect of the treatment on the response. ▪ Experiments that are randomized occur when the subjects are randomly assigned to the treatments; randomization helps to eliminate the effects of lurking variables. 4 Copyright © 2013, 2009, and 2007, Pearson Education, Inc. 3 Components of a Good Experiment 1. Control/Comparison Group: allows the researcher to analyze the effectiveness of the primary treatment. 2. Randomization: eliminates possible researcher bias, balances the comparison groups on known as well as on lurking variables. 3. Replication: allows us to attribute observed effects to the treatments rather than ordinary variability. 5 Copyright © 2013, 2009, and 2007, Pearson Education, Inc. Component 1: Control or Comparison Group ▪ A placebo is a dummy treatment, i.e. sugar pill. Many subjects respond favorably to any treatment, even a placebo. ▪ A control group typically receives a placebo.
    [Show full text]
  • Partial Confounding Balanced and Unbalanced Partially Confounded Design Example 1
    Chapter 10 Partial Confounding The objective of confounding is to mix the less important treatment combinations with the block effect differences so that higher accuracy can be provided to the other important treatment comparisons. When such mixing of treatment contrasts and block differences is done in all the replicates, then it is termed as total confounding. On the other hand, when the treatment contrast is not confounded in all the replicates but only in some of the replicates, then it is said to be partially confounded with the blocks. It is also possible that one treatment combination is confounded in some of the replicates and another treatment combination is confounded in other replicates which are different from the earlier replicates. So the treatment combinations confounded in some of the replicates and unconfounded in other replicates. So the treatment combinations are said to be partially confounded. The partially confounded contrasts are estimated only from those replicates where it is not confounded. Since the variance of the contrast estimator is inversely proportional to the number of replicates in which they are estimable, so some factors on which information is available from all the replicates are more accurately determined. Balanced and unbalanced partially confounded design If all the effects of a certain order are confounded with incomplete block differences in equal number of replicates in a design, then the design is said to be balanced partially confounded design. If all the effects of a certain order are confounded an unequal number of times in a design, then the design is said to be unbalanced partially confounded design.
    [Show full text]
  • Types of Confounding
    Confounding What is confounding? Why we need it? Advantage and Disadvantage. How to reduce block size? Types of Confounding. Examples. ANOVA TABLE Concept of Confounding. In a factorial experiment we have n factors each are at s levels. If the factors and level of the factors are increased, then total no. of treatment combinations will also increase which we call the effect of heterogeneity that is the variability among the treatment combination will increase and consequently the errors variance will increase. Because of this result the estimate will be least precise estimate. So one is interested to remove the variability among the treatment combination. which is only possible if the size of the block will reduce to make homogeneity in the block. This concept of reducing the block size is call confounding. Who has developed confounding theory? • R.A. Fisher • Yates Definition Confounding of factorial experiment is defined as reduction of block size in such a way that one block is divided into two or more blocks such that treatment comparison of that main or interaction effect is mixed up with block effect. Main or Interaction effect which is mixed up with block effect are called Confounded effect and this experiment is called Confounding experiment. Block Block-1 Block-2 n0p0 n1p0 n0p0 n1p0 n1p1 n0p1 n0p1 n1p1+n1p0 n0p1+n0p0 n1p1 • Block Effect :- (n1p1+n1p0 ) – (n0p1+n0p0 ) • Treatment Effect.:- =Main effect N = (n1-n0) (p1+p0) =n1p1 +n1p0-n0p1-n0p0 =(n1p1+n1p0 ) – (n0p1+n0p0 ) Thus, Treatment comparison – Block effect =0 Here, We can say that main effect N is mixed up with block effect .
    [Show full text]
  • Experimental Design
    A few remarks on Experimental Design Simon Anders • How many replicates do I need? • What is a suitable replicate? • What effects can I hope to see? A preliminary: the word “sample” Statistician: “Our sample comprises 20 male adult subjects chosen at random from the patient population.” Biologist: “For our experiment, we used blood samples from 20 patients chosen at random.” An observed effect is called statistically significant. What does this mean? p < 0.05 ?? If the experiment were repeated with new, independently obtained samples, the effect would likely be observed again. If the experiment were repeated with new, independently obtained samples, the effect would likely be observed again. When can we claim such a thing? Only if we tried more than once! Replication How many replicates? What is the average height of a man in Germany? each dot: height of one subject each dot: height of one subject What is the average height of a man in Germany? each dot: average of n subjects standard deviation of observations standard error of mean = sample size Purposes of replication: 1. More replicates allow for more precise estimates of effect sizes. each dot: height of one subject What is the average height of a man in Germany? each dot: average of n subjects each dot: standard deviation, estimated from n subjects standard deviation of observations standard error of mean = sample size Purposes of replication: 1. More replicates allow for more precise estimates of effect sizes. 2. Replicates allow to estimate how precise our effect-size estimates are. Two Martian scientists, Dr Nullor and Dr Alton, ask: Is body height correlated with hair colour in human males within a city’s population? Dr Nullor’s UFO observes Helsinki: µ(dark) = µ(blond) = 180 cm.
    [Show full text]
  • Agricultural Statistics, Why Replication and Randomization Are Important Authors: Robert Mullen
    4/26/2016 C.O.R.N. Newsletter 2005­40 | Agronomic Crops Network Agricultural Statistics, Why Replication and Randomization are Important Authors: Robert Mullen In CORN Newsletter 2005­39 (http://corn.osu.edu/story.php?issueID=117&layout=1&storyID=657), we had a lengthy discussion as to why replication and randomization are necessary for statistical analysis. From a theoretical perspective the article was important, but from an applied perspective why do we care? Today we will discuss the potential influence of factors other than those we are trying to evaluate on experimental outcomes, and how proper statistical design can ensure the conclusions reached are correct. Replication Assume you want to evaluate a fungicide treatment on your farm, so you split a field in two and apply the treatment to one half and leave the other half untreated. At the end of the year you harvest each of the two halves and observe a 3 bushel per acre increase in yield on the treated side. This 3 bushel per acre difference seems like a good deal, so you decide that next year all of your acres will be treated with this new fungicide. Are you sure that the additional 3 bushels per acre was due to the application of the fungicide? Closer inspection of the field reveals that the half of the field that showed the yield response was dominated by a lighter texture soil that drained better than the other half of the field. Due to excessive moisture (this is not necessarily this past year) the half of the field with better drainage might be expected to perform better.
    [Show full text]
  • Aspects of Design for Repeated Measurements R
    17 ASPECTS OF DESIGN FOR REPEATED MEASUREMENTS R. MEAD The design of trials and investigations is extremely important and includes the choices of experimental or observational units, the form and timing of measurements, the choice of applied or environmental treat­ ments and the interrelationship of these three components. It became obvious later during the workshop that many difficulties of analysis could be attributed to design deficiencies (in a broad sense). The general principles of design2 applicable to all investigations are: (a) efficient use of resources (b) asking many questions (c) reducing a 2 • The consequent statistical principles are: (i) Using appropriate amounts and forms of replication. (ii) Using random allocation of treatments to units within stated de­ sign restriction to provide a valid basis for inferences and for esti­ mating a 2 • (iii) Using small "blocks" to control random variation. (iv) Using factorial structure with the implied priority ordering of ef­ fe,cts. (v) Using the minimum necessary number of levels of quantitative factors. The aspects discussed in this paper are (i), (iii) and (v). 1. REPLICATION AND RESOURCES Replication for comparative mean values must provide adequate power; for random variance a 2 and a difference that should be detected, 18 d, this implies a replication, n, where Note that n may include implicit, as well as explicit, replication, derived from treatment structure or regression. For a given replication level there will be a total amount of re­ sources, represented by the total degrees of freedom (N -1) in an anal­ ysis of variance. These resources are used in three ways: (i) Asking treatment questions (ii) Estimating 0'2 (iii) Controlling variation.
    [Show full text]