The first and most crucial question that confronts any author is "who am I writing for?" The audience to a large extent determines both the style and content of any written work.
You should assume that your book review will be read by an intelligent and literate reader, but one who may not be well versed in the minutiae of your particular subject. This means, of course, that you will have to define any uncommon terms, explain important events, and avoid the use of jargon. Do not, however, carry your exposition too far and bore your reader with detailed summaries and/or unimportant trivia.
You should begin with an examination of the author's thesis (by thesis, I mean the author's main argument, contention, or idea). Every author is trying to convince the reader of something. What is it? An author may be arguing not only for a particular thesis, but also against some thesis(es) that preceded his/her work. In this case, you should note not only the author's thesis but also the criticisms being made of earlier works. This is especially so when an author is presenting a new thesis that is opposed to an older and more generally accepted one.
After you have discovered the author's thesis, you must move on and examine the evidence he/she musters to support the argument. Here you are the jury listening to an attorney (the author) presenting his arguments (the thesis). You must look at the evidence and decide if it is indeed compelling and sufficient. What type of sources (primary or secondary) does the author use? Here footnotes and the bibliography (usually at the end of the book) are of great help. Does the author use original source documents (primary material) or does he/she rely on the work of previous historians (secondary material)? Do the sources cited appear biased to you? If one was writing a biography of Eli Whitney, for instance, and used only one-sided sources such as the reminiscences of Whitney's son, the work should be held suspect. What are the argument's greatest strengths and weaknesses? Can you think of other arguments or sources that might possibly strengthen (or weaken) the thesis? Are you convinced?
A good book review also will pay some attention to the author's writing style. Does the author possess a good command of the English language? Does the book move along in a fluid manner, or is it jumpy with poor transitions from section to section? Does the author blend the thesis into the general narrative and avoid repetition? What is your overall assessment of the book? Did you learn from it, even enjoy it? Would you recommend it to others?
One great aid in learning how to prepare a critical book review is to read several reviews that other scholars have written about the work. All of the major historical journals, such as the American Historical Review and the Journal of American History, have in each issue a section devoted to book reviews. So do more specialized journals, such as Technology and Culture, Isis, Labor History, and the Business History Review. Read several reviews of the book you have selected just to see what the reviewers have attempted to do and how they have done it. This will doubtless help you in preparing your own review. I would suggest, however, that you do this after you have read the book, taken notes, and formed your own conclusions. Don't be afraid to give your candid opinion; you do not have to agree with what other reviewers have said.
Another word of caution. Generally book reviewers have a pet thesis or two of their own, and these almost always show up in their reviews. Imagine, for example, a General Motors executive reviewing a Ralph Nader report on automobile safety. Bear this very important fact in mind, not only as you read book reviews, but also as you write your own.
How do you find book reviews in scholarly journals? One option is to consult JSTOR (which is provided as a site-license subscription at many universities). Another way is to go to the library and look up the review in a journal like Technology and Culture or the Journal of American History. Most books are reviewed within a year or two of publication. Suppose, for example, the book you read was published in 1985 and you want to see if it was reviewed in the journal, Technology and Culture (the primary journal for the history of technology). All you need to do is to go to the stacks and check the index of the journal for the years 1985 through 1987 or 1988 under the author's name. The index will tell you in which issue the review appeared. Another admittedly old fashioned way of locating reviews of books is to check the appropriate entry (vis. the author's name) in the Book Review Digest and that will lead you to reviews that may have appeared in well-known publications like the New York Times, the American Historical Review, Reviews in American History, and the like.